53877 WO R L D B A N K WO R K I N G PA P E R N O. 1 8 1 Restoring Balance Bangladesh's Rural Energy Realities M. Asaduzzaman Douglas F. Barnes Shahidur R. Khandker THE WORLD BANK W O R L D B A N K W O R K I N G P A P E R N O . 1 8 1 Restoring Balance Bangladesh's Rural Energy Realities M. Asaduzzaman Douglas F. Barnes Shahidur R. Khandker THE WORLD BANK Energy Sector Management Assistance Program Washington, D.C. Copyright © 2010 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: March 2010 Printed on recycled paper 1 2 3 4 13 12 11 10 World Bank Working Papers are published to communicate the results of the Bank's work to the development community with the least possible delay. The manuscript of this paper therefore has not been prepared in accordance with the procedures appropriate to formally-edited texts. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the International Bank for Reconstruction and Development/The World Bank and its affiliated organizations, or those of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank of the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission promptly to reproduce portions of the work. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, Tel: 978-750-8400, Fax: 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, Fax: 202-522-2422, email: pubrights@worldbank.org. ISBN-13: 978-0-8213-7897-7 eISBN: 978-0-8213-8310-0 ISSN: 1726-5878 DOI: 10.1596/978-0-8213-7897-7 Library of Congress Cataloging-in-Publication Data has been requested. Contents Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Acronyms and Abbreviations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Units of Measure/Currency Equivalents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Energy Conversion Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 What Is the Knowledge Base? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Study Goal and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Principles for a Rural Energy Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Study Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Survey Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Structure of This Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2. Household Energy Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 General Consumption Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Energy for Cooking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Energy for Lighting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Summing Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3. Household Energy Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Expenditure Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Cooking Energy: Biomass and Its Opportunity Cost. . . . . . . . . . . . . . . . . . . . . . . . . 21 Lighting Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Asset Ownership Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Demand for Other Energy: Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Summing Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 4. Household Gains from Energy Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Biomass Cooking: From Traditional to Modern Energy . . . . . . . . . . . . . . . . . . . . . . 33 From Kerosene to Electric Lighting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 From Manual to Mechanized Farming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Overall Income Gains: Moving Up the Energy Ladder . . . . . . . . . . . . . . . . . . . . . . . 44 Summing Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 iii iv Contents 5. Energy Consumption and Rural Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Enterprise Types and Distribution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Growth-Center Microenterprises. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Home and Village Enterprises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Enterprise Energy Demand and Profitability: A Quantitative Analysis . . . . . . . . . . 62 Energy for Farming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Summing Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 6. Institutional Framework for Rural Energy Supply . . . . . . . . . . . . . . . . . . . . . . . . . 69 Institutional Challenges and Potential. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Building an Integrated Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 7. Policy Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Problems and Potential of Biomass Energy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Rural Electrification and Rural Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 Toward Policies for Household Petroleum Fuels . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Enhancing Rural Energy Projects, Policies, and Strategies . . . . . . . . . . . . . . . . . . . . 91 The Way Forward. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 ANNEXES 1 Selected Tables from the Household Survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 2 Statistical Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 3 Consumer's Surplus as a Measure of Welfare . . . . . . . . . . . . . . . . . . . . . . . . . . 133 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 List of Special Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 LIST OF BOXES 1.1 What Are the Benefits of Rural Energy?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.1 Modern Energy Benefits for Rural Families . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 LIST OF FIGURES ES.1 Energy Use and Expenditures in Rural Bangladesh, 2004 . . . . . . . . . . . . . . . xxiii 1.1 Types and Sources of Energy Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.2 Map of Survey Areas in Bangladesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.1 Rural Household Consumption by Source (percent energy consumption). . . . . 14 3.1 Energy Expenditure Distribution by Energy Source (percent expenditure) . . . . 20 3.2 Household Biomass Collection Time by Income Quintile (thousands of Tk per year). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Contents v 3.3 Regional Variation in Electrification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 4.1 PM10 Concentration by Time of Day and Household Location . . . . . . . . . . . . 35 4.2 PM10 Concentration by Age Group and Gender . . . . . . . . . . . . . . . . . . . . . . . . 36 4.3 Pollution Factors and PM10 Concentrations. . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 5.1 Comparison of Energy Use Structure by Enterprise Location . . . . . . . . . . . . . 60 6.1 Annual Load Shedding, 1993­2004. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 6.2 Bangladesh Power Sector Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 6.3 Oil and Gas Sector Organization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 A3.1 Consumer Demand Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 A3.2 Consumer's Surplus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 LIST OF TABLES Cooking Fuels and Energy Efficiencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix Household Lighting Energy in Developing Countries . . . . . . . . . . . . . . . . . . . xx Battery Electricity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xx 1.1 Number of Entities in the Census and Sample by Type and Division . . . . . . . . 8 1.2 Coverage of Growth Centers by Division . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.1 Household Distribution of Energy Sources by Division (percent) . . . . . . . . . 12 2.2 Annual Household Consumption (physical quantity) . . . . . . . . . . . . . . . . . . . 13 2.3 Energy Sources for Cooking and Parboiling of Rice by Division (annual kgoe per household) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.4 Distribution of Biomass Supply Sources by Energy Type . . . . . . . . . . . . . . . . . 15 2.5 Energy Sources for Lighting (annual kgoe per household). . . . . . . . . . . . . . . . 16 2.6 Electrification of Villages and Households by Division . . . . . . . . . . . . . . . . . . 17 3.1 Annual Energy Expenditure by Income and Division (Tk per household). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 3.2 Biomass Collection Time by Income and Division (annual hours per household) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.3 Household Electrification Connection Rate and Consumption by Income Quintile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.4 Comparison of Kerosene and Electric Lighting Services. . . . . . . . . . . . . . . . . . 24 3.5 Annual Price of Lighting by Income Quintile and Division . . . . . . . . . . . . . . . 25 3.6 Price of Electric Lighting by Income, Connection Type, and Division . . . . . . 26 3.7 Household Ownership of Cooking Stoves by Type (number per 100 households) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.8 Average Effective Cooking Hours by Income Quintile . . . . . . . . . . . . . . . . . . . 27 3.9 Household Ownership of Lighting Appliances by Electrification Status (number per 100 households) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 vi Contents 3.10 Electric Lighting by Income and Division . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.11 Ownership of Non-lighting Electric Appliances by Income Quintile . . . . . . . 30 3.12 Summary Statistics of Outcomes and Important Explanatory Variables of Household Energy Demand Regressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 3.13 Estimates of Household Energy Demand. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 4.1 Mean Daily Hours in Household Location by Age Group . . . . . . . . . . . . . . . . 34 4.2 Annual Energy Use by Number of Students in Household . . . . . . . . . . . . . . . 40 4.3 Study Time by Household Electrification Status (average number of hours) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 4.4 Households by Lighting System and Income . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 4.5 Household Expenditure by Electrification Status on Kerosene and Other Energy Sources, by Income Group (Tk per year) . . . . . . . . . . . . . . . . . . . . . . . 41 4.6 Welfare Gains from Switching Lighting Systems . . . . . . . . . . . . . . . . . . . . . . . . 42 4.7 Farm Household Distribution by Tillage/Irrigation Technologies . . . . . . . . . 43 4.8 Gross Productivity of Paddy by Tillage/Irrigation (Tk/decimal) . . . . . . . . . . . 43 4.9 Paddy Productivity in Boro Season by Land Ownership (gross value in Tk/decimal) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4.10 Household Income Differentials by Electrification Status (Tk per year) . . . . . . 46 4.11 Summary Statistics of Outcomes and Energy Use Variables of Household Welfare Regressions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 4.12 Estimates of Household Welfare (Tk per month) . . . . . . . . . . . . . . . . . . . . . . . 47 5.1 Microenterprise Distribution by Location and Activity Type . . . . . . . . . . . . . 52 5.2 Distribution of Growth-Center Microenterprises by Region . . . . . . . . . . . . . 53 5.3 Distribution of Energy Sources by Activity Type . . . . . . . . . . . . . . . . . . . . . . . . 54 5.4 Major Energy Sources by Use Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 5.5 Energy as a Percentage of Total Costs by Activity Type . . . . . . . . . . . . . . . . . . . 55 5.6 Price per Unit of Energy for Non-lighting Business Use. . . . . . . . . . . . . . . . . . 56 5.7 Perceived Advantages and Disadvantages of Electricity Use . . . . . . . . . . . . . . . 57 5.8 Perceived Advantages and Disadvantages of Fuelwood Use . . . . . . . . . . . . . . . 58 5.9 Perceived Advantages and Disadvantages of Kerosene Use . . . . . . . . . . . . . . . 58 5.10 Microenterprise Distribution by Location and Length of Operation . . . . . . . 59 5.11 Microenterprise Distribution by Number of Employees and Location. . . . . . 60 5.12 Comparison of Energy Output and Cost by Use and Location . . . . . . . . . . . . 61 5.13 Summary Statistics of Outcomes and Explanatory Variables of Enterprise Energy Demand Regressions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 5.14 Estimates of Enterprise Energy Demand (N = 2,290) . . . . . . . . . . . . . . . . . . . . 63 5.15 Summary Statistics of Outcomes and Energy Use Variables of Enterprise Profitability Regressions (N = 2,290) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Contents vii 5.16 Effects of 1 Tk Increased Energy Use on Enterprise Outcome (annual Tk per enterprise) (N = 2,290) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 5.17 Comparison of Tillage Methods by Division . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 5.18 Irrigation Method and Energy Type by Division . . . . . . . . . . . . . . . . . . . . . . . . 66 5.19 Farmer Irrigation Method by Division . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 5.20 Energy Consumption in Farming (average owner-user households per year) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 A1.1 Survey Sample (number) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 A1.2 Household Distribution by Ownership of Agricultural Assets (percent households) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 A1.3 Household Distribution by Various Characteristics (percent distribution). . . . . 99 A1.4 Annual Household Income by Source (average Tk per household). . . . . . . . 100 A1.5 Household Distribution by Income Source (percent). . . . . . . . . . . . . . . . . . . 100 A1.6 Household Consumption Expenditure by Budget Item (Tk per year) . . . . . 101 A1.7 Household Distribution by Type of Energy Consumption (percent) . . . . . . 101 A1.8 Annual Household Energy Consumption: All Divisions (average per household) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 A1.8.1 Annual Household Energy Consumption: Dhaka (average per household) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 A1.8.2 Annual Household Energy Consumption: Rajshahi (average per household) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 A1.8.3 Annual Household Energy Consumption: Chittagong (average per household) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 A1.8.4 Annual Household Energy Consumption: Khulna (average per household) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 A1.9 Average Energy Consumption for Agriculture and Transportation (owner-users only) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 A1.10 Energy Consumption by Business Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 A1.11 Household Distribution of Cooking Stove Ownership (percent) . . . . . . . . . 106 A1.12 Household Ownership of Cooking Stoves by Type (number per 100 households) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 A1.13 Household Distribution of Number of Meals and Cooking Times (percent distribution). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 A1.14 Hours of Operating Cooking Stoves by Type (daily). . . . . . . . . . . . . . . . . . . . 107 A1.15 Household Ownership of Lighting Appliances (number per 100 households) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 A1.16 Hours of Running Lighting Appliances (cases per day) . . . . . . . . . . . . . . . . . 108 A1.17 Household Ownership of Plug-in Electric Appliances (number per 100 households) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 A1.18 Hours of Running the Plug-in Type of Electric Appliances (daily) . . . . . . . . 108 viii Contents A1.19 Ownership of Irrigation Pumps and Power Tillers/Tractors (percent households) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 A1.20 Hours of Running Pumps and Power Tillers/Tractors (annual hours per unit) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 A1.21 Cost of Household Energy Consumption: All Divisions (Tk per household per year) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 A1.21.1 Cost of Household Energy Consumption: Dhaka (Tk per household per year) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 A1.21.2 Cost of Household Energy Consumption: Rajshahi (Tk per household per year) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 A1.21.3 Cost of Household Energy Consumption: Chittagong (Tk per household per year) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 A1.21.4 Cost of Household Energy Consumption: Khulna (Tk per household per year) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 A1.22 Household Biomass Energy Behavior: All Divisions (percent distribution). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 A1.22.1 Household Biomass Energy Behavior: Dhaka (percent distribution) . . . . . . 112 A1.22.2 Household Biomass Energy Behavior: Rajshahi (percent distribution) . . . . 112 A1.22.3 Household Biomass Energy Behavior: Chittagong (percent distribution). . . . . 113 A1.22.4 Household Biomass Energy Behavior: Khulna (percent distribution) . . . . . 113 A1.23 Household Consumption of Selected Biomass Energies: All Divisions (per year) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 A1.23.1 Household Consumption of Selected Biomass Energies: Dhaka (per year). . . . 114 A1.23.2 Household Consumption of Selected Biomass Energies: Rajshahi (per year) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 A1.23.3 Household Consumption of Selected Biomass Energies: Chittagong (per year) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 A1.23.4 Household Consumption of Selected Biomass Energies: Khulna (per year) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 A1.24 Operational Land and Household Ownership of Gardens . . . . . . . . . . . . . . . 115 A1.25 Household Ownership of Trees and Bamboo . . . . . . . . . . . . . . . . . . . . . . . . . 115 A1.26 Household Ownership of Cattle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 A1.27 Household Distribution by Duration of Electricity Connection (percent distribution). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 A1.28 Electricity Sources (percent of connected households). . . . . . . . . . . . . . . . . . 116 A1.29 Basis of Charging for Electricity (percent of connected households) . . . . . . 117 A1.30 Household Distribution by Electricity Supply Problem (percent of connected households) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 A1.31 Cost of Kerosene--Using Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 A1.32 Cost of Electric Bulbs and Tube Lights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Contents ix A1.33 Women's Time Use for Various Activities (daily hours per person). . . . . . . . 118 A1.34 Time Spent Making Clay Stoves (hours per stove) . . . . . . . . . . . . . . . . . . . . . 118 A1.35 Time Spent on Clay Stove Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 A1.36 Respondents' Attitude toward Electricity (percent distribution) . . . . . . . . . . 119 A1.37 Respondents' Attitude toward Fuelwood (percent distribution) . . . . . . . . . . 120 A1.38 Respondents' Attitude toward Kerosene (percent distribution) . . . . . . . . . . . 120 A1.39 Respondents' Attitude toward LPG/LNG (percent distribution) . . . . . . . . . . 120 A1.40 Respondents' Knowledge of Price of Kerosene (percent distribution) . . . . . 121 A1.41 Respondents' Reasons for Costliness of Kerosene (percent of those who reported kerosene as costly). . . . . . . . . . . . . . . . . . . . . 121 A1.42 Respondents' Knowledge of Price of Electricity (percent distribution). . . . . 121 A1.43 Respondents' Reasons for Costliness of Electricity (percent of those who reported electricity as costly) . . . . . . . . . . . . . . . . . . . . 121 A1.44 Respondents' Knowledge of Renewable Energy (percent distribution). . . . . 122 A1.45 Respondents' Reasons for Interest in Purchasing Renewable Energy Devices (percent of those interested in purchasing). . . . . . . . . . . . . . 122 A1.46 Respondents' Reasons for Disinterest in Purchasing Renewable Energy Devices (percent of those not interested in purchasing) . . . . . . . . . . 123 A2.1 Tobit Estimate of Household Energy Demand (N = 2,388) . . . . . . . . . . . . . . 125 A2.2 Effects of Energy Use on Household Outcomes by Use Type (N = 2,388) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 A2.3 Effects of Energy Use on Household Outcomes by Energy Source (N = 2,388) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 A2.4 Tobit Estimate of Enterprise Energy Demand (N = 2,290). . . . . . . . . . . . . . . 130 A2.5 Effects of Enterprise's Total Energy Use on Profitability (N = 2,290) . . . . . . 131 A2.6 Effects of Enterprise's Energy Use on Profitability by Source (N = 2,290) . . . . . 131 Acknowledgments T his study was conducted with technical and financial support from the Energy Sector Management Assistance Program (ESMAP), the collaborative efforts of the Bangladesh Institute of Development Studies (BIDS) in Dhaka and the World Bank. In addition, a detailed background survey of microenterprises was completed by Data International, Dhaka. The survey team completed its work in 2005 and the consultants submitted a preliminary report to the World Bank in 2006. The World Bank team completed the revised report by 2007 and after extensive peer review by the South Asia Regional Energy Program the report was finalized and submitted to ESMAP for publication in late 2008. At that time, the report was put into production under the guidance of ESMAP. It is the intention of the authors that you find this report interesting and insightful. The principal project leaders were: Douglas F. Barnes and Shahidur R. Khandker, who supervised the international World Bank team comprised of: M. Asaduzzaman and Abdul Latif who led the local BIDS team, and Najmul Hossain who was the Data Inter- national project manager. Special thanks go to Keith Openshaw and Rashid Faruqee, Hussain Samad, and Grayson Heffner of the World Bank team for having made extremely valuable contributions to the project, including the drafting of sections of this report. Thanks also go to M. Iqbal, who coordinated all work in Dhaka as part of the World Bank team. The work on institutional issues was originally prepared by M. Khalequzzaman, who also commented extensively on the final draft of the report. An early review of rural energy issues in Bangladesh, prepared by M. Nurul Islam, was quite useful in setting the overall context of rural energy in Bangladesh. Preparation of this report would not have been possible without the dedicated efforts of the BIDS staff and consultants. We greatly appreciate the comments of the World Bank peer review team, which vastly improved the final version of the report. Peer reviewers included: Priti Kumar, Dana Rysankova, Chrisantha Ratnayake, Raihan Elahi, Sudeshna Bannerji, Sameer Akbar, and Marlon Lezama. Salman Zaheer, Manager of the South Asia Energy Unit, guided the review meeting for the report and also made many valuable comments on the report. We also thank the staff of the Ministry of Power, Energy, and Mineral Resources, whose construc- tive comments over the course of the project and on the final results were useful in com- pleting this report. Finally, we thank former and current staff members of the Energy Sector Management Assistance Program, for their help in making this report possible. Ede Jorge Ijjasz-Vasquez provided support for the completion of this work. Also, Dominique Lallement provided both earlier assistance and direction in completing this report. Marjorie Araya from ESMAP and Thaisa Tiglao from ESMAP (now EXTOP) played a crucial role in assisting throughout all stages of the publication process, including production of the study in its final form. Finally, the project owes a debt of gratitude to Norma Adams, who edited the final version of this report and played a major role in enhancing its quality. xi Acronyms and Abbreviations ARI acute respiratory infection BBS Bangladesh Bureau of Statistics BCSIR Bangladesh Council for Scientific and Industrial Research BIDS Bangladesh Institute of Development Studies BPC Bangladesh Petroleum Corporation BPDB Bangladesh Power Development Board CNG compressed natural gas ERC Energy Regulatory Commission ESMAP Energy Sector Management Assistance Program EXTOP World Bank Office of the Publisher FAO Food and Agricultural Organization GOB Government of Bangladesh HIES Household Income and Expenditure Survey IAP indoor air pollution IDCOL Infrastructure Development Company Limited IEA International Energy Agency IFRD Institute of Fuel Research and Development LNG liquefied natural gas LPG liquefied petroleum gas NGO Non-governmental organization PBS Rural Electric Cooperative (Palli Bidyut Samity) PGCB Power Grid Company of Bangladesh PM10 particles with an aerodynamic diameter less than 10 microns PM2.5 particles with an aerodynamic diameter less than 2.5 microns PV photovoltaic REB Rural Electrification Board SEDA Sustainable Energy Development Agency SHS solar home systems xiii Units of Measure GJ gigajoule GWh gigawatt hour kg kilogram kgoe kilograms of oil equivalent klm-hr kilolumen-hour kWh kilowatt-hour loe liters of oil equivalent MCF thousand cubic feet MJ megajoule MW megawatt µg/m3 micrograms per cubic meter W watt Currency Equivalents 2004 US $1 = 60 Bangladesh takas 2005 US $1 = 64 Bangladesh takas 2006 US $1 = 68 Bangladesh takas 2007 US $1 = 69 Bangladesh takas 2008 US $1 = 68 Bangladesh takas 2009 US $1 = 69 Bangladesh takas xv Energy Conversion Factors Cooking Fuels and Energy Efficiencies Energy content Cooking stove Efficient cooking Fuel source (MJ per unit) efficiency (%) energy (MJ per unit) Electricity (kWh) 3.6 70 2.5 LPG (kg) 45.5 60 27.3 Natural gas (m3) 38.0 60 22.8 Biogas (60% methane) (m3) 23.0 60 13.8 Kerosene, pressure (kg) 43.0 55 23.7 Kerosene, wick (kg) 43.0 35 15.1 Charcoal, efficient (kg) 30.0 30 9.0 Charcoal, traditional (kg) 22.5 20 6.0 Bituminous coal (kg) 22.5 25 5.6 Fuelwood, efficient (15% moisture) (kg) 16.0 25 4.0 Fuelwood, traditional (15% moisture) (kg) 16.0 15 2.4 Crop residues (5% moisture) (kg) 13.5 12 1.6 Dung (15% moisture) (kg) 14.5 12 1.7 Leaves and grass (kg) 13.5 12 1.6 Source: O'Sullivan and Barnes (2006) and various World Bank reports. Note: Energy content = total energy by energy source; efficient cooking energy = energy absorbed into cooking pans or devices. xvii xviii Energy Conversion Factors Household Lighting Energy in Developing Countries Output cased on lumens per watt Lighting type Light output (lumens) klm-hr per kgoe klm-hr per kWh Non-electric Paraffin candle 11.8 2.33 0.20 Kerosene, wick 11.4 1.15 0.10 Kerosene, hurricane 32 1.92 0.16 Kerosene, pressure 2,040 17.53 1.48 Incandescent (watts) 25 230 109.12 9.20 40 430 127.50 10.75 50 580 137.58 11.60 60 730 144.30 12.17 100 1,280 151.80 12.80 Florescent (watts) 10 600 711.63 60.00 20 1,200 711.63 60.00 40 1,613 478.27 40.33 Compact florescent Philips lamp (15-W) 894 706.88 59.60 Philips lamp (9-W) 369 486.28 41.00 Osram sol lamp (6.14-W) 240 463.60 39.09 Source: O'Sullivan and Barnes (2006). Battery Electricity Storage capacity Usable storage capacity Lighting type Power rating (amps) (watt hrs.) (watt hrs.) D Cell (high-quality battery) 0.5 6 4.8 Chinese D Cell 0.2 2 1.6 Car battery (60 amp) 60.0 720 576.0 Source: O'Sullivan and Barnes (2006). Note: Battery amp hours * 12 = watt hours capacity. 20 percent is lost in charging. A battery can only discharge to 80 percent of capacity (amp hours*12*.8). Executive Summary S ince the 1990s, Bangladesh--one of the world's poorest countries--has taken large strides toward achieving the Millennium Development Goals (MDGs), having outperformed most low-income countries on a range of social indicators. The country is on target to achieve a two-thirds reduction in infant and child mortality by 2015. Food security has improved significantly, and gender disparity in primary and secondary enrollment has been eliminated. Yet in 2005, nearly 40 percent of the nation's 140 million residents and 44 percent of rural residents were below the poverty line (World Bank 2006).1 Bangladesh is one of the world's lowest energy producers. Electricity, critical to economic growth, has reached only about one-third of households, despite the country's successful rural electrification program. With the exception of kerosene, commercial fuels are beyond reach for many. Biomass fuels, collected mainly from the local environment only two decades ago, are fast becoming a marketed commodity as access to local biomass becomes ever more difficult. The stark reality is that many rural residents are dependent on such fuels as agricul- tural residues, dung, and even leaves and grass for cooking. Knowledge of rural energy in Bangladesh lags recent developments in energy technolo- gies and policies. Until recently, the potential for renewable energy and the adverse health effects of indoor air pollution (IAP) on women and children were ignored. Research on the structure and efficiency of rural energy markets has also been overlooked. Policy analysis and recommendations for rural energy development, along with their implications for the institutions that implement them, have not been assigned a high priority. With the exception of rural electrification, most intervention programs have empha- sized commercial energy, consumed mainly in urban areas. At the same time, the rural elec- trification program has continued to expand, a forestry master plan has been approved for implementation, and both the government and donors are supporting efforts to popular- ize improved biomass stoves. Yet the effectiveness of these initiatives, in terms of people's overall energy-using behavior, remains under the radar screen of many development researchers. Study Goal Lack of comprehensive data and analysis and knowledge gaps regarding Bangladesh's current rural energy reality have precluded the development of a strategy for modern and efficient energy use. Against this backdrop, the World Bank Group, through a consulta- tive process, initiated two surveys--one focused on rural households and the other on village microenterprises and rural growth centers--to elicit information on energy-using behavior and characteristics. Subsequently, the Bank commissioned studies on the mar- ket structure for energy and the macro-level dimensions of biomass supply and demand. This study--the first to concentrate on Bangladesh's energy systems and their effects on the lives of rural people--drew on these background studies, as well as other World Bank­financed research on IAP and rural infrastructure, to present a rural energy strategy for the country. The study's broad aim was to identify ways to improve the living standard xix xx Executive Summary in rural Bangladesh through better and more efficient use of energy, while creating an environment conducive to growth and poverty reduction. For any developing country, the crux of a rural energy strategy is to have more and better choices for meeting rural demand for energy through market mechanisms and sound policy. This goes hand in hand with the development of competent implementing institu- tions, which are critical to the process. Also important are new supply- and demand-side technologies that can be used to raise rural people's welfare and improve productivity to increase growth prospects. Accordingly, the rural energy strategy advocated by this study aims to satisfy the types of demand that increase household welfare and raise rural growth prospects as energy becomes a direct input into the production process. Benefits of Moving up the Energy Ladder For Bangladesh to grow and prosper, its rural economy must not be ignored. Focusing solely on urban growth--representative of only about 20 percent of the country's 140 mil- lion people--would lead to inequitable social and economic development. Making the benefits of modern energy services available in rural areas can promote decentralized development and growth and help rural residents become more productive, thereby mit- igating urban­rural disparities. Cooking Energy Biomass continues to play a critical role in Bangladesh's rural energy balance; today, it is just as important, if not more so, than 25 years ago. Unlike some other South Asian coun- tries, where liquefied petroleum gas (LPG) and other modern fuels have begun to enter the marketplace, rural Bangladesh continues to depend heavily on biomass fuels (fuel- wood, cow dung, crop residues, and tree leaves and grass) to meet household cooking needs (see Figure 1). Some 95 percent of Bangladeshi households collect or purchase biomass energy with which to cook all or part of their meals, mainly using fixed clay stoves. The inherent inef- ficiency of such stoves, combined with the high moisture content of biomass cooking fuels, results in incomplete combustion, producing IAP. In rural Bangladesh, age-sex composi- tion of households indicates that generally, women and their children suffer greatest expo- sure. Women also bear the brunt of time spent collecting biomass fuels, accounting for an average of 150 out of 200 household hours per year. Switching to improved biomass cookstoves and modern cooking fuels can enable rural families, especially women, to improve their quality of life. Because improved stoves require about 25 percent less fuel than traditional ones, using them can similarly reduce the amount of time spent collecting biomass fuels or the money used to purchase them. Cook- ing with modern petroleum fuels, such as LPG, entails similar welfare gains. In addition to time and cost savings, cooking with cleaner fuels avoids the health risks associated with IAP. Lighting Energy In rural Bangladesh, some 70 percent of lighting energy is derived from kerosene, while electricity accounts for most of the other 30 percent. Households with electricity prefer Executive Summary xxi Figure ES.1. Energy Use and Expenditures in Rural Bangladesh, 2004 % Average Energy Use % Average Energy Expenditures Electricity (Including Imputed Collection Values) Kerosene 1% Others Others Non-grid 2% 0.3% 1% electricity Grid 3% electricity 10% Dung 17% Firewood Firewood Kerosene 44% 38% 12% Crop residue Dung 21% 14% Crop Tree Tree residue leaves leaves 13% 9% 15% Source: BIDS Survey, 2004. electric lighting to kerosene, but unreliable supply drives them to use kerosene lamps as a backup source in case of power failure. In terms of cost and kilolumen output, electricity is far superior to kerosene. Electricity's output per unit of energy is 100 times higher than that of kerosene. When one compares households that use kerosene lamps or electric lighting, two trends are obvious. First, the level of lighting increases with income for both kerosene and electricity. Second, the quality of lighting service obtained from electricity is an order of magnitude higher than that from kerosene for all income groups across all regions. For the average household, the number of kilolumens is nearly 80 times higher for those who use mainly electricity, versus kerosene. The high-quality lighting afforded by electricity has important consequences for Bangladeshi household welfare. A household's electricity consumption for lighting appears to rise steadily as the number of students in the household increases. The study's findings show that total hours of study time are greater in households that use electric lighting than in those that use kerosene lamps, confirming electricity's contribution to human-capital for- mation. Apart from education, household income also shows a distinct trend in lighting energy use; as incomes rise, households increasingly switch from kerosene- to electricity- based lighting. The potential gain from this transition, as measured by consumer's surplus, is enormous--40 to 45 percent of household income. A separate quantitative analysis shows net positive gains in household income from lighting energy use of similar levels, confirm- ing that lighting has a high value for rural households. Farming Energy A third critical energy use for rural Bangladeshi households is farming. The country's potential for increased irrigation, though not as developed as that of other South Asian countries, is significant. The most obvious energy inputs are diesel engines and electric xxii Executive Summary motors for pumping irrigation water. In addition, various small machines, including power tillers, are used increasingly in land preparation, alleviating much of the drudgery associ- ated with such work. This study concluded that farm households that move up the energy ladder enjoy a substantial gain in productivity. For households that use mechanized instead of manual irrigation, farm productivity is 15 to 23 percent higher. Econometric analysis shows that farm income increases as much as 6.3 percent from diesel use and 7.9 percent from electricity use in irrigation. The study's village survey found that 68 percent of villages used mechanized irrigation methods; of these, 70 percent used diesel power, 10 percent used electricity, and 3 percent used both. One possible reason for farmers not having adopted electric pumps to the extent that they have diesel ones is that 34 percent of villages remain without electricity. More- over, even in villages with electricity, the potential gain of electricity (7.9 percent) over diesel (6.3 percent) may not suffice to warrant large-scale substitution; that is, both diesel and electric pumps result in similar levels of improved farm productivity. But electricity can be used in tube wells, a common practice in other South Asian countries; while diesel engines are generally limited to surface wells. Thus, in rural Bangladesh, the unrealized potential application of electricity is widespread, presumably for reasons of policy. Microenterprise Energy and Productivity Switching to modern energy and the modern use of biomass also contributes to the produc- tivity of rural microenterprises, representing a substantial increase in rural residents' income. The three types of enterprises examined in this study--growth-center, village, and home- based--use a wide array of energy sources. All rural businesses in the growth centers use elec- tricity, either from the grid or local generators; and most have electric lighting. By contrast, not all free-standing village and home-based enterprises have access to electricity; those that do may choose not to use it. Such enterprises depend heavily on biomass energy, mainly for heating and manufacturing needs. Though inconvenient, biomass is the least expensive heat- ing fuel; also, most rural areas lack access to LPG. Increasingly, Bangladeshi entrepreneurs perceive the health problems associated with traditional fuelwood use, as well as the growing scarcity of local biomass. They also view traditional lighting sources as inefficient and of poor quality, but unreliability of electric- ity supply forces businesses to depend on kerosene and diesel as backup lighting sources. Analysis of the relationship between modern energy sources, as defined by the energy ladder concept, indicates that rural businesses that use modern energy generally have more revenue and are more profitable than those that rely on traditional energy sources. Thus, a viable rural energy strategy for Bangladesh must consider the best ways to pro- mote modern energy, as well as the modern use of biomass. Energy Access In rural Bangladesh, access to energy is governed by its availability, pricing (both monetary and non-monetary), household income, and other characteristics. One major reason for the prominence of kerosene for lighting is that electricity is not available in all areas. Even where it is available, consumer density is low. Thus, while 66 percent of the villages sam- Executive Summary xxiii pled in this study had a grid connection, only 29 percent of households were connected. Even in the capital division of Dhaka, where one would expect a higher density, less than 50 percent of households adopted electricity, although 80 percent of villages had some type of electricity service available. If electricity is available, household access is determined by income: the higher the income (approximated by land and non-land assets), the greater the demand. With higher incomes, rural Bangladeshis were found to pursue improved sources of energy and invest in assets that require more energy. Higher-income rural households typically diversify their energy portfolios, moving toward modern energy, which is not only more efficient but also allows household members to invest time and money in more productive activities than biomass energy alone would make possible. For any energy source, pricing also determines demand. Thus, energy access is not equitable because the source is either priced too high or carries a high opportunity cost (e.g., biomass collection). Inequitable access and level of use may lead to further inequities (e.g., low-income households may pay far higher prices for equivalent lighting services). Paying More Attention to Rural Energy Services Rural energy must be viewed as a basic input to the rural economy, in line with its role in rural productivity and income generation. Achieving a supply-and-demand balance is critical at all levels. Given rural residents' heavy reliance on traditional biomass sup- ply, biomass must be used in more efficient ways that mitigate damage to human and global environmental health. Complementing these efforts, more diversified modern energy sources must be made available to fulfill unmet rural household and business demand. To this end, appropriate pricing policies are vital: they must be market-based to ensure that suppliers can sustain the higher cost of rural operations, yet remain afford- able to the poor. Biomass: Increasing Use Efficiency and Supply Mitigating the ill effects of biomass burning requires both demand- and supply-side solu- tions adapted to local realities. Biomass demand can be reduced by adopting improved cooking stoves that rural people want and harnessing biogas technology through village- based production and distribution networks. Biomass supply can be increased by planting trees around homes, maintaining and improving local natural resources, and increasing agricultural productivity. Promoting a sustainable supply-demand balance implies the need for policy initiatives that encourage intermediate and end-use efficiency, especially at the household level; increased productivity to keep pace with growing demand; and fuel substitution, where possible. Rural Electrification: Toward a Pro-poor Approach Both grid and off-grid electrification programs are critical to the socioeconomic develop- ment of rural Bangladesh. As this study shows, electrification translates into substantial gains in household welfare and a higher quality of life. As electricity demand is projected xxiv Executive Summary to double over the next nine years, the investment required to generate sufficient capacity to accommodate future demand growth is immense. If electricity is available, households will pursue its use. But nationwide access to grid electricity is only 40 percent; in rural areas, it is just 30 percent. The Rural Electri- fication Board (REB), responsible for electrifying rural Bangladesh, is based on the U.S. model of consumer-owned rural electric cooperatives. Today, rural Bangladesh is divided into 70 cooperatives or Palli Bidyut Samities (PBSs), benefiting an estimated 7 million households--a remarkable achievement in a country of some 22 million rural households and about 110 million rural residents. Although the PBSs are generally well managed, the program has a nationwide electricity reliability problem. Some 80 percent of rural households report daily outages, while 60 percent report significant power fluc- tuations. Survey results bring into question the REB's selected coverage approach, sug- gesting a too-stringent connection policy or other barriers that prohibit many poor households from receiving electricity directly. In addition, most of the PBSs are not yet financially viable. Given that only 3 percent of Bangladesh's 22 million rural households are gaining access to the national grid each year, it could take several decades to reach the entire rural popula- tion because, as the electricity grid expands, it becomes increasingly expensive to serve remote households. Thus, the government and policy makers, along with international donor sup- port, should continue and even increase the successful expansion program financed through the REB. Concurrently, off-grid systems can complement this program to meet the enormous unmet need for electricity. In 2002, the Infrastructure Development Company Limited (IDCOL) was made the country's focal agency for coordinating the off-grid program, the first phase of which has focused on promoting solar photovoltaic (PV) systems. To date, most work has been done by non-governmental organizations (NGOs) specialized in micro- finance and microenterprise development. Working with 16 partner organizations, includ- ing Grameen Shakti and other NGOs involved in microcredit and solar home systems (SHS), the program has succeeded in installing more than 80,000 systems over a three-year period. It is also important to diversify product lines to micro-grids, which could be connected directly to the national grid system once the PBSs reach more remote communities. Household Petroleum Fuels: Toward Equitable Use For the past 30 years, the Bangladesh Petroleum Corporation (BPC) and its subsidiaries have controlled most aspects of petroleum supply, including its pricing system. Kerosene is priced uniformly across the country, but transport-cost adjustments are made for market distances greater than 40 kilometers from a supply depot. The price difference between rural and urban areas is not great, and the system for containing kerosene within a price range works well throughout the country. By contrast, LPG, which is widely available in urban and peri- urban markets, has largely failed to reach rural households. By switching from biomass to LPG to meet a portion of their cooking needs, rural family members, especially women, could realize large time savings and better health. Finally, piped natural gas (methane) is used mainly by householders in Dhaka and other large towns. But under current pricing policy, households are charged a (subsidized) flat monthly rate, irrespective of the amount of gas consumed. Such a policy leads to abuse and waste and discriminates against house- holds without connections in smaller towns and rural areas. Executive Summary xxv Importance of Sound Pricing and Subsidy Policies Promoting more equitable rural access and use of modern energy implies the need for sound pricing policies to increase supply and reliability. It should be noted that the urban areas of Bangladesh are served by a state utility, which is generally considered to be less effi- cient than the rural cooperatives electricity companies. Even though performing better, the rural electricity policies and service still can be improved. For rural grid electricity, a care- ful review of the policies of uniform subsidies and similar tariffs for all PBSs is recom- mended. One potential solution might include mechanisms that give preferential treatment to companies serving poorer regions, such as permitting higher tariffs or providing elec- tricity at lower bulk prices. To date, off-grid electricity schemes have succeeded by provid- ing subsidies through individual projects, allowing prices to float based on the cost of service after subsidies.2 As the previous discussion suggests, pricing policies for household petroleum fuels also require careful review. Kerosene, used by most rural households for lighting, is accessible to most consumers at world-market rates, and is available in quantities sufficient for cooking. Thus, its pricing policies require no substantial revision. For LPG, it is recommended that the upfront costs of needed stoves and cylinders be partially subsidized or paid for in install- ments, which would put LPG within reach of households that otherwise could not afford it. Developing and making available smaller LPG cylinders is occurring in Bangladesh, and should be encouraged even further. This would help extend the reach of LPG into less wealthy markets. In addition, a level playing field should be promoted for all public and pri- vate companies that market LPG to encourage innovation and expansion into more remote, underserved areas. With regard to natural gas, it is recommended that the government reverse its flat-rate tariff so that households are charged per unit of consumption and that subsidies, which invariably benefit wealthier households, be removed. Institutional Challenge: The Way Forward Rural energy is a complex issue, encompassing a broad and diverse spectrum of resources-- from household petroleum fuels to biomass and renewable energy--spanning multiple sectors, including forestry, electricity, and health. To date, the many diverse institutions addressing rural energy issues in Bangladesh have been poorly coordinated. One major rec- ommendation resulting from this study is to develop the long-term institutional capacity to tackle rural energy issues in all their complexity. Such an institution would promote rural energy solutions through its technical assistance, advice to government, and facilita- tion of grants and loans for worthy projects. Rural energy's importance to the Bangladesh economy cannot be underestimated, given the world's focus on globalization and market reform. The problems rural people face in obtaining safe, clean, and reliable energy supplies are not minor inconveniences. On the contrary, they represent a significant barrier to rural economic development and improved social well being. A multifaceted approach to solving Bangladesh's rural energy problems is an essential building block to propel the country into the twenty-first century. The rural energy strategy identified is this: Bangladesh has a comprehensive need for better institutional coordination and attainment of a critical mass of technology and market development. With effective institutional coordination, combined with market xxvi Executive Summary development, appropriate subsidy, pricing policies, and government and donor support, current and proposed programs can succeed beyond expectations. The call for action is urgent, not only for rural development, but for the country's equitable and overall eco- nomic growth. Notes 1. See Bangladesh Country Brief (www.worldbank.org). 2. Through the World Bank­financed Rural Electrification and Renewable Energy Development Project (RERED), IDCOL has worked with selected local partners to provide rural consumers a 20 percent subsidy for SHS using a microfinancing scheme. CHAPTE R 1 Introduction B angladesh is one of the world's poorest countries.1 Nearly 80 percent of the nation's 140 million people reside in rural areas; of these, 20 percent live in extreme poverty. Geographically, many low-lying areas are vulnerable to severe flooding, while other regions are prone to drought, erosion, and soil salinity. Such an unfavorable agricultural landscape, combined with mismanagement of natural resources and increasing population pressure, is pushing many of the rural poor to the brink. Because Bangladesh is such a poor country, it also is one of the world's lowest energy producers. Total annual energy supply is only about 150 liters of oil equivalent per capita (IEA 2003); in rural areas, conditions are even worse. Compared to other developing coun- tries, Bangladesh uses little modern energy. Despite its successful rural electrification pro- gram, close to two-thirds of households remain without electricity and, with the exception of kerosene, commercial fuels are beyond reach for many. Moreover, biomass fuels are becoming increasingly scarce. Collected mainly from the local environment as recently as two decades ago, biofuels are fast becoming a marketed commodity as access to local biomass continues to shrink.2 Today, people in rural Bangladesh use a mosaic of energy sources to meet their various domestic and productivity needs. Biomass is still used extensively for cooking, while kerosene is the main lighting source. Electrified households prefer electricity for both lighting and appliance use. Many small businesses, including home-based enterprises, use electricity for lighting, agricultural processing, and other productive activities. Gradually, the types of energy used in rural areas are improving. Diesel-powered irriga- tion pumps and agricultural tillers are more common than in the past, when manual and ani- mal powers were the traditional energy sources for cultivation. Grid-based electrification is reaching more rural households at a faster pace. Recent projects in renewable technologies, including solar home systems (SHS), reflect slow, yet significant, progress. In sum, while the 1 2 World Bank Working Paper traditional biomass system continues to predominate, rural Bangladesh is moving slowly, yet inexorably, toward modern forms of energy and more efficient use of biomass as options for energy supply. What Is the Knowledge Base? In the mid-1970s, Bangladesh conducted its first large-scale energy study, which was fol- lowed in the early 1980s by a more comprehensive study that included a rural energy com- ponent (GOB 1987). Now, more than two decades later, little new information, with the exception of small rural energy surveys, is available. Indeed, knowledge of rural energy in Bangladesh lags recent developments in energy technologies and policies. Studies on rural electrification and development have been conducted over the past decade; but many new issues remain unexplored. Until recently, the potential for renew- able energy and the adverse health effects of indoor air pollution (IAP) on women and children were ignored. Research on the structure and efficiency of rural energy markets has also been overlooked. Policy analysis and recommendations for rural energy devel- opment, along with their implications for the institutions that implement them, have not been assigned a high priority. With the exception of rural electrification, most intervention programs have empha- sized commercial energy, consumed mainly in urban areas. At the same time, the rural elec- trification program has continued to expand.3 A forestry master plan has been approved for implementation, and various community and agroforestry projects have been implemented. The country has renewed its emphasis on tree planting and management as part of activities to mitigate climate change. To curb the health hazards of IAP, both the government and donors are supporting efforts to popularize improved biomass stoves. Yet the effectiveness of these initiatives, in terms of people's overall energy-using behavior, remains under the radar screen of many development researchers. One of the country's major economic policies is to promote agriculture and rural devel- opment. Indeed, food production and consumption are inextricably linked to rural energy. But this reality--recognizing the energy needs for various rural economic and development activities--appears as an afterthought in many government documents. Such policy docu- ments equate energy with electricity, gas, and hydrocarbons.4 But the main energy uses in rural areas involve fuelwood, cow dung, and crop residue; rural Bangladeshis use these fuels in large quantities to cook their food, heat their homes (in cooler regions), produce bricks in kilns, and make molasses (gur) from sugar cane. Study Goal and Objectives Lack of comprehensive data and analysis and knowledge gaps regarding Bangladesh's cur- rent rural energy status have precluded the development of a strategy for modern and effi- cient energy use. Against this backdrop, the World Bank, through a consultative process, initiated two surveys--one focused on rural households and the other on village micro- enterprises and rural growth centers--to elicit information on energy-using behavior and Restoring Balance: Bangladesh's Rural Energy Realities 3 characteristics. Subsequently, the Bank commissioned studies on the market structure for energy and the macro-level dimensions of biomass supply and demand (Openshaw 2004; Khalequzzaman 2005). This study--the first to concentrate on Bangladesh's energy systems and their effects on the lives of rural people--drew on these background studies, as well as other World Bank­financed research on IAP and rural infrastructure, to present a rural energy strat- egy for the country. Much of this study's analytical underpinning was based on several background studies. This study also reanalyzed data from earlier research to better under- stand the benefits of modern energy use for rural households, farm activities, and small businesses. This study's broad aim was to identify ways to raise the living standard in rural Bangladesh via better and more efficient energy use, while creating an environment conducive to growth and poverty reduction. To this end, the study sought to do the following: Describe the rural energy situation for households and small businesses. Determine the demand for and benefits of moving up the energy ladder. Describe the energy-supply situation and identify bottlenecks to improved supply. Recommend policies to help rural regions move toward improved productivity and higher quality of life. Based on these analyses, the study put forward a set of recommendations as part of a rural energy strategy for Bangladesh. Principles for a Rural Energy Strategy The rural energy strategy of any developing country must be guided by basic principles. First, because the types of energy services that people demand may be obtained from more than one source, issues of cost-effectiveness and satisfying people's preferences arise; not all services may be equally desirable because of their price or aesthetic or practical value. Second, because any particular fuel may have a variety of uses, issues of fuel substitution and complementarity also are important. Household demand for energy primarily involves cooking, lighting, heating, productive uses, and other end uses. Rural enterprises have a significant demand for non-residential process heat. Both households and rural enterprises require lighting, space cooling and heat- ing, and motive power (among productive activities, crop cultivation generates demand for motive power for irrigation pumps, tillers, and transport) as illustrated in Figure 1.1. The crux of a rural energy strategy is to have more and better choices for meeting rural demand for energy through the market mechanism and sound policy. This goes hand in hand with the development of competent implementing institutions, which are critical to the process. Also important are new supply- and demand-side technologies that can be used to raise rural people's welfare and improve productivity to increase growth prospects. Accordingly, the rural energy strategy advocated by this study aims to satisfy the types of demand that increase household welfare and raise rural growth prospects as energy becomes a direct input into the production process (see Box 1.1). 4 World Bank Working Paper Figure 1.1. Types and Sources of Energy Demand Home, village, Tree biomass conventional forest Domestic cooking, parboiling Animal dung Livestock Agricultural Agriculture Residues Bio-gas Lighting Domestic mines ENERGY DEMAND Kerosene Nonresidential, Import process heat Coal Space heating or Gas, LPG, CNG Domestic gas fields cooling Import Petroleum fuels Motive power/traction/ transport/pump Grid, PVs, batteries Electricity generator Amusement/leisure Draught power Source: Bangladesh Institute of Development Studies (BIDS) (2005). Study Method This study was based mainly on the results of two large-scale primary surveys, one on rural households and the other on rural growth centers and microenterprises defined in the Bangladesh context as large marketplaces. Because of their importance, these surveys, along with background information, are described in greater detail in this chapter.5 Furthermore, selected information from the surveys was reanalyzed to address the welfare and growth effects of modern energy use. Restoring Balance: Bangladesh's Rural Energy Realities 5 Box 1.1. What Are the Benefits of Rural Energy? For Bangladesh to grow and prosper, its rural economy must not be ignored. Focusing solely on urban growth--representative of only about 20 percent of the country's 140 million people-- would lead to inequitable social and economic development. Making modern energy services and jobs available in rural areas can help rural residents become more productive, thereby mitigating urban­rural disparities and reducing rural-to-urban migration. The benefits of rural energy range from increased time savings and farm productivity to improved education, communication, and overall quality of life. For example, electric lights emit more than 100 times the amount of lighting provided by a kerosene lantern or candle, permitting household members--both adults and children--to read and study during evening hours. Electric irrigation pumps make it possible to grow and market crops year-round and raise rural incomes. Electric lighting and motive power can increase the productivity of women- owned microenterprises. Availability of radios and televisions can improve communication and information sharing. Use of well-designed stoves and efficient cooking fuels can reduce IAP and the number of hours that families, especially women and children, spend collecting biomass. Taken together, these benefits can lead to greater rural productivity and a higher quality of life for many of the country's poorest citizens. Source: Barnes and Floor (1996). The study also drew on reports prepared specifically to assess Bangladesh's biomass and forestry situation, relevant institutions and policies, and delivery mechanisms for sev- eral fuel types in rural areas. Various other secondary materials were used, including an earlier World Bank report on IAP in Bangladesh (Dasgupta et al. 2004). Survey Instruments This study was based on information from a comprehensive survey, whose objectives were to determine the overall energy-use patterns, potential market for modern energy in rural areas, willingness to pay for electricity services, and barriers to adopting renewable energy.6 To meet these objectives, and thereby assist in the development of a rural energy strategy, these surveys were conducted: Household survey. This study determined the socioeconomic characteristics, energy demand and availability, consumer ability and willingness to pay, attitudes toward various energy sources, and perceived benefits of energy. Growth center and microenterprise survey. This study developed village-level profiles, including characteristics and potential energy demand, and assessed the energy used by small- and medium-sized businesses located in village marketplaces, known as growth centers. Because 25 years had passed since Bangladesh had conducted such a major rural energy survey, the present one used representative samples of rural households and commercial businesses (see Table A1.1, Annex 1). 6 World Bank Working Paper Household Survey The household survey, conducted in 2004, was based on a cluster sampling strategy. Nationwide, a total of 40 rural subdistricts (known as upazilas or thanas) were randomly selected from four older divisions (proportionate to population). This was followed by the random selection of three villages from each subdistrict (see Figure 1.2). In order to develop a village sample frame, a census was conducted in each village. This census covered all households; village enterprises based outside the home but not located at the growth center; and all educational, religious, and health institutions. Samples were taken of households, enterprises, and institutions outside the home for further in-depth data collection (see Table 1.1). In addition, a community survey was conducted to identify village characteristics. The resulting sample closely resembled Bangladesh's overall rural population. The survey revealed that about 88 percent of households are headed by men 26 to 45 years of age, nearly 50 percent of whom are illiterate. The average household size is 5. Nearly 55 percent own little or no land (0.5 acres maximum); only 6 to 7 percent own more than 5 acres. About 30 percent of households own little more than their homestead. In 2004, annual household income was roughly US$1,000 (Tk 62,000) or US$200 per per- son. Some 43 percent of this amount was derived from agricultural sources (e.g., crop cul- tivation, non-crop agriculture, and agricultural wages), while the remainder was from non-agricultural sources (e.g., processing and trading activities and remittances). All of these figures closely correspond to equivalent figures for rural Bangladesh. In the case of village-based enterprises, several had more than one distinct business activity because they were owned or operated by households that, when interviewed for the household survey, were found to have home-based enterprises. These 137 enterprises were treated separately. These rural energy survey results compare favorably with those of the Bangladesh Household Income and Expenditure Survey (HIES), conducted in 2005. For both surveys, the percentage of rural households with electricity was close to 30 percent. There were slight differences in fuel expenditures. According to this rural energy survey, the average annual expenditure for grid electricity was Tk 488 and Tk 608 for kerosene. Results from the 2005 HIES showed that rural households' average annual expenditure for grid elec- tricity was Tk 522 and Tk 566 for kerosene. The results are close, given the different meth- ods used to measure these expenditures. These findings indicate that rural households have increased their electricity consumption while lowering their dependency on kerosene. Growth Center and Microenterprise Survey Because most of the country's rural microenterprises are located in growth centers, a separate survey of microenterprises in rural growth centers was conducted to obtain a representative sample of enterprises located outside the home. Given that no prior information existed on enterprise distribution in the growth cen- ters, a two-phase random sampling technique was adopted. The first phase involved the random selection of 40 subdistricts (from a list of 460 subdistricts, covering six divisions), from which three growth centers (per subdistrict) were chosen. The second phase involved a sample survey of business establishments. From each growth center, 15 or more enter- prises were randomly selected. The first step in the selection process entailed a census of Restoring Balance: Bangladesh's Rural Energy Realities 7 Figure 1.2. Map of Survey Areas in Bangladesh Source: This map was prepared by the World Bank Map Design Unit (2008). Shaded areas are sample upazilas (political units) from the Bangladesh Rural Energy Survey (BIDS, 2005). 8 World Bank Working Paper Table 1.1. Number of Entities in the Census and Sample by Type and Division Census Type Chittagong Dhaka Khulna Rajshahi Total Household 5,178 5,928 5,508 5,611 22,225 Business enterprise 342 487 659 394 1,882 Institution 201 223 249 184 857 Total 5,721 6,638 6,416 6,189 24,964 Survey type Household 640 603 548 600 2,391 Home enterprise 31 50 29 27 137 Other enterprise 63 85 136 58 342 Institution 36 38 50 32 156 Source: Asaduzzaman and Latif (2005). the enterprises, which provided the sample frame for enterprise selection in each growth center. Because of the small number of growth centers in some subdistricts, the number of growth centers sampled was 115 instead of the 120 targeted for the study. As a result, for several locations, more than the minimum required sample was covered. The total num- ber of enterprises studied from the 115 growth centers was 1,801. Dhaka, Chittagong, and Rajshahi accounted for about 75 percent of regional distribution; while Khulna, Sylhet, and Barisal constituted the remainder (see Table 1.2). Implications for a Rural Energy Strategy The two surveys described here represent the first comprehensive attempt to assess the rural energy needs of and benefits for Bangladesh's rural households and microenterprises. Along with a variety of other materials, the surveys provide the analytical underpinnings of this study--the first in 25 years to examine Bangladesh's complex rural energy situation. By covering both households and microenterprises, the study has allowed us to examine rural energy-use patterns and the adequacy of the country's current rural energy policies Table 1.2. Coverage of Growth Centers by Division Growth Centers Shops Division Total No. Sampled Total in Growth Centers No. Sampled Dhaka 531 29 101,256 477 Chittagong 373 25 64,022 403 Rajshahi 543 31 91,592 456 Khulna 274 15 41,685 226 Sylhet 157 3 20,986 48 Barisal 180 12 34,635 191 Bangladesh 2,058 115 354,175 1,801 Source: Data International (2004). Restoring Balance: Bangladesh's Rural Energy Realities 9 and institutional framework. The study moves toward broad recommendations that can support a higher quality of life for rural residents and, thus, more equitable economic growth for the country. Structure of This Report The structure of this report reflects the directional organization of the study. Chapter 2 presents an overview of the rural energy situation of Bangladeshi households, moving up the energy ladder--from biomass cooking energy to kerosene and electric lighting. Chapter 3 offers an in-depth analysis of household energy demand, examining expendi- ture patterns for cooking, lighting, and appliance ownership. It presents the results of an econometric analysis to predict how household energy demand is affected by changes in such factors as household income and energy pricing. Chapter 4 provides a descriptive analysis of the household welfare gains that can result from transitioning up the energy ladder, including better health from use of improved biomass stoves and improved edu- cation from electric lighting. For farm households, making the switch from manual to diesel- and electric-powered equipment is examined in terms of gains in agricultural pro- ductivity. Chapter 5 explores the effects of modern energy on rural production, focusing on the energy-using characteristics of microenterprises and farming. Chapter 6 turns to the current institutional framework for rural energy supply and implications for meeting rural energy demand. It focuses mainly on delivery mechanisms of the biomass, power, oil and gas, and renewable energy (non-biomass) sectors. Finally, Chapter 7 suggests a broad institutional strategy to meet the complex challenges of Bangladesh's current and future rural energy needs. Notes 1. The World Bank estimates that, in 2005, 28 percent of the urban population and 44 percent of rural residents were below the poverty line, with a nationwide poverty rate of 40 percent (World Bank 2006). 2. In 2005, the FAO Forestry Department reported that only 6.69 percent of Bangladesh's total land area (13.02 million hectares) remained under forest cover. 3. A major study has been conducted on the effects of rural electrification on development (Barkat et al. 2002); at the time of this writing, a monitoring and evaluation study, involving a large survey, was under way. 4. In various government documents, the energy chapter discusses electric power, gas, coal, and lique- fied petroleum gas (LPG); while rural energy is relegated to the forestry section of the agriculture chapter. The National Energy Policy (GOB 1996), a decade-old document, reflects little on the importance of rural energy. Similarly, the government's recent poverty reduction strategy makes little mention of rural energy (GOB 2005). 5. For the basic technical analysis of these surveys, see Asaduzzaman and Latif (2005). 6. A portable document format (PDF) file of survey modules and questionnaires is available from ESMAP upon request. CHAPTE R 2 Household Energy Use U nlike South Asian regions where liquefied petroleum gas (LPG) and other mod- ern fuels have entered the marketplace, rural Bangladesh still depends heavily on fuelwood--as well as dung, crop residue, and even tree leaves--for cooking fuel. Indeed, in many areas of the country, biomass energy is in short supply. At the same time, the importance of kerosene, electricity, and LPG is growing. This chapter provides a snap- shot of rural Bangladeshi households' energy-use situation, focusing on general con- sumption patterns and those for the key energy uses of cooking and lighting. General Consumption Patterns Nearly all households use both biomass and non-biomass energy (see Table 2.1). A vast majority use biomass sources, including fuelwood, tree leaves, and crop residue. While kerosene is the predominant non-biomass energy source, many households also use grid electricity and dry-cell batteries. Biomass is used almost exclusively for cooking (see Table 2.2). Fuelwood constitutes 41 percent of total biomass cooking energy. Surprisingly, 39 percent is represented by agri- cultural residue or leaves and grass, indicating a shortage of fuelwood energy.1 Kerosene is used primarily for lighting; grid electricity is another important lighting source. Monthly average consumption of kerosene is 2.25 liters, which barely covers basic lighting services, even considering that households with access to grid electricity use little kerosene for lighting. Batteries, an expensive but extensively used source of electricity, are In addition to the main authors, M. Abdul Latif was a contributing coauthor of this chapter. 11 12 World Bank Working Paper Table 2.1. Household Distribution of Energy Sources by Division (percent) Division Energy Source Chittagong Dhaka Khulna Rajshahi All Divisions Biomass 98.6 99.8 100.0 99.5 99.5 Fuelwood 95.8 85.1 88.7 67.2 84.3 Tree leaves 61.6 81.4 91.8 72.0 76.1 Crop residue 53.6 81.4 75.4 93.2 75.5 Dung cake/stick 29.5 56.9 64.8 72.3 55.2 Sawdust 0.3 0.7 1.6 0.8 0.8 Non-biomass 100.0 100.0 100.0 96.5 99.1 Kerosene 98.4 99.0 100.0 91.5 97.2 Grid electricity 38.9 43.8 10.8 20.3 29.0 Dry-cell battery 42.3 39.1 61.3 50.7 48.0 Candle 10.6 2.5 4.4 0.8 4.7 LPG/LNG 1.1 -- -- 0.2 0.3 Natural gas 0.9 -- -- -- 0.3 Storage cell 0.8 1.3 0.9 -- 0.8 Solar PV -- 0.3 1.5 -- 0.4 Source: BIDS Survey (2004). convenient for powering flashlights, radios, and other consumer electronics. Grid electricity is mostly derived from the cooperatives or PBSs (Palli Bidyut Samities), which serve rural areas through the national grid system. Nearly 30 percent of rural households use electric- ity provided by the PBSs, and the annual percentage continues to climb. This figure com- pares well with that of the Bangladesh Household Income and Expenditure Survey (HIES), conducted during the same period, which was 30 percent. Fuelwood is also the single most important rural energy source in terms of energy unit (kgoe), accounting for some 44 percent of total consumption (see Figure 2.1). Including tree leaves and twigs, the share of tree-based biomass is nearly 60 percent of total house- hold energy. Crop residue (e.g., bagasse, jute sticks, rice hulls, bran; as well as various types of straw and uprooted plant remains) and animal residue (e.g., cow dung made into round cakes or sticks and dried before burning) constitute other major sources. Not generally used for cooking, modern fuels account for only 3 percent of the energy balance. Because of the significant amount of energy required by cooking and the inefficiency of most rural stoves, the useful or delivered energy is lower than the percentages presented; however, they highlight the importance of biomass for cooking. Estimated annual energy consumption by rural households is 1,049 kgoe or 8.9 giga- joules (GJ) per person--a vast increase from the 5 GJ per person consumed 25 years ago. Over the same period, the annual growth rate was more than 2.6 percent, higher than the average growth rate in per capita income. Indeed, when only biofuels are considered, growth appears even more remarkable. Leach (1987), using figures based on Islam (1980, 1986), reported that rural households used an estimated 4.2 GJ of biofuels. The corresponding fig- Restoring Balance: Bangladesh's Rural Energy Realities 13 Table 2.2. Annual Household Consumption (physical quantity) Energy Use Heating Energy Source Cooking Parboiling Other Cooling Lighting Amusement All Uses Biomass (kg) Fuelwood 1,064.84 28.60 92.77 -- -- -- 1,186.21 Tree leaves 470.67 29.99 0.85 -- -- -- 501.51 Crop residue 538.86 164.41 2.72 -- -- -- 708.18 Dung cake/ 503.68 16.07 4.16 -- -- -- 523.90 stick Sawdust 8.36 0.02 0.02 -- -- -- 8.40 Non-biomass Kerosene 1.76 -- 0.07 -- 27.16 -- 28.98 (liter) Grid 0.25 -- 4.00 49.50 80.74 9.34 143.83 electricity (kWh) Dry-cell -- -- -- -- -- -- 15.01 battery (piece) Candle -- -- -- -- 15.86 -- 15.86 (piece) LPG/LNG 0.05 -- -- -- -- -- 0.05 (liter) Natural 9.59 -- -- -- -- -- 9.59 gas (Tk) Storage -- -- -- -- 0.14 0.41 0.55 cell (kWh) Source: BIDS Survey (2004). ure for biomass in the current study's household survey is 8.6 GJ per person; this figure translates into an average annual growth rate of 3.2 percent, outstripping even population growth rate.2 This finding highlights the critical role that biomass continues to play in the rural energy balance; today it is just as important, if not more so, than 25 years ago. Energy for Cooking The energy required for basic household cooking--including the parboiling of rice-- mirrors, in large part, the country's overall rural energy situation; yet one observes major regional differences (see Table 2.3). For example, in Khulna and Chittagong, regions home to major forests, consumption levels are higher. Indeed, nearly 67 percent of Chittagong's cooking energy is derived from fuelwood, indicating that plentiful availability leads to greater use. Not surprisingly, Rajshahi, the division least endowed with forest resources, 14 World Bank Working Paper depends least on tree-based bio- Figure 2.1. Rural Household Consumption by Source mass (fuelwood and leaves) for (percent energy consumption) household cooking. In Khulna, where one might expect higher Kerosene Electricity Other consumption levels, strict forest 2% 1% 0% regulations and restricted access Dung keep consumption lower than in 17% Firewood Chittagong, where local residents 44% can more easily access scattered forest patches. Crop Significant use of biomass residue 21% energy--especially tree leaves, crop residue, and dung--means Tree leaves that indoor air pollution (IAP) 15% and local biomass shortages Source: BIDS Survey (2004). are potential problems in rural Note: The "Electricity" category includes only grid-based areas of all four divisions. Since electricity. Other types of electricity (solar PV, storage fuelwood is a comparatively cells, and dry-cell batteries) were either nearly absent or more efficient fuel with a higher could not be quantified; whatever could be quantified is energy content, the potential for included in the "Other" category. IAP may be somewhat less in Chittagong. The little use made of improved biomass stoves that vent smoke to the outdoors and modern fuels are important related issues, which are discussed in Chapters 3 and 4. Commercialization of Biomass Energy A major finding of this study, as mentioned previously, is the increasing commercialization of biomass energy in rural areas. Just two decades ago, most biomass fuels were collected from the local environment. Fuelwood, conventionally derived from rural residents' own production or local collection, is today more likely to be purchased from local markets. Table 2.3. Energy Sources for Cooking and Parboiling of Rice by Division (annual kgoe per household) Chittagong Dhaka Khulna Rajshahi All Divisions Energy Source C P C P C P C P C P Fuelwood 644.8 7.0 365.5 16.7 404.6 9.7 180.5 9.70 400.4 10.7 Tree leaves 120.3 2.6 157.6 12.1 218.0 19.3 116.2 5.40 149.7 9.5 Crop residue 117.4 27.0 159.4 36.9 183.1 114.7 269.4 37.70 171.3 52.3 Dung 79.4 0.8 155.8 7.6 220.6 4.5 251.4 9.20 171.7 5.5 Sawdust 1.4 0 1.0 0 5.8 0 2.8 0.02 2.6 0.01 Kerosene 1.8 0 2.8 0 0.6 0 0.6 0.00 1.4 0 Total 965.0 37.4 842.1 73.4 1,032.7 148.2 820.9 62.00 897.3 78.1 Source: BIDS Survey (2004). Note: C = cooking; P = parboiling. Restoring Balance: Bangladesh's Rural Energy Realities 15 Table 2.4. Distribution of Biomass Supply Sources by Energy Type Energy Type (%) Supply Source Fuelwood Tree Leaves Crop Residue Dung Cake/Stick Chittagong Own production 14.0 57.8 69.4 82.8 Gathered 61.9 41.2 28.8 3.6 Purchased 24.1 1.1 1.8 13.6 Dhaka Own production 6.3 55.5 72.2 23.2 Gathered 26.9 43.9 15.0 9.2 Purchased 66.8 0.6 12.8 67.6 Khulna Own production 7.6 55.5 66.6 92.5 Gathered 32.4 40.3 25.7 1.3 Purchased 60.0 4.2 7.8 6.2 Rajshahi Own production 7.3 41.0 67.5 25.0 Gathered 55.4 58.9 28.3 17.1 Purchased 37.3 0.1 4.2 57.9 All divisions Own production 11.0 52.2 68.5 72.2 Gathered 49.8 46.3 24.7 5.3 Purchased 39.3 1.4 6.8 22.5 Source: BIDS Survey (2004). Indeed, about 40 percent of all fuelwood is now purchased from local markets; in Dhaka, the proportion is nearly 67 percent (see Table 2.4). Given that fuelwood constitutes the major portion of energy consumption, its increasing commoditization has major implications for energy costs, an issue addressed in Chapter 3. The marketplace is also a major source of other biomass energy supplies, including dung cakes or sticks, particularly in Dhaka and Rajshahi. As Table 2.4 indicates, more than half of tree leaves are still gathered, while most crop residue is derived from residents' own production. Recent development of rural markets for biomass energy reflects an increase in over- all supply; but perhaps more important, a decrease in local supply. The implications for cooking fuels and their increasing scarcity in the local environment suggest the need for policy dialogue on the evolving status of cooking fuels in rural Bangladesh. Energy for Lighting In rural Bangladesh, some 70 percent of energy consumed for lighting is derived from kerosene and most of the other 30 percent from electricity (Box 2.1). Nearly 100 percent of 16 World Bank Working Paper Box 2.1. Modern Energy Benefits for Rural Families Making the switch to electric lighting and appliances, petroleum fuels, and improved cooking stoves can enable rural families to raise their incomes and improve their quality of life. In a hot tropical climate such as Bangladesh, the addition of a simple electric fan can significantly improve a rural household's indoor comfort level and ward off insects. Electric lighting offers 100 times more light than traditional kupi or kerosene lamps commonly used in households without electricity. The higher quality of lighting makes it possible for families to pursue read- ing and other educational activities during evening hours. Cooking with kerosene or LPG, or using improved biomass stoves--still a rare occurrence in rural Bangladesh--can result in fewer hours spent collecting biofuels, less cooking time, and reduced IAP. Chapter 4 considers these benefits in more detail. Source: Barnes and Floor (1996). households claim to use kerosene lamps. Although households with electricity prefer electric lighting to kerosene, unreliable supply drives them to use kerosene lamps as a backup in case of power failure. While overall energy consumption for lighting is low, regional differences are substantial. For example, rural households in Chittagong use more than twice as much energy as those in Rajshahi (because of differences in kerosene and electricity consumption). In Khulna, nearly 92 percent of lighting is derived from kerosene (see Table 2.5). A major reason for the prominence of kerosene is that electricity is not available in all areas; even where available, consumer density is low. Thus, while 66 percent of the villages sampled had a grid connection, only 29 percent of households were connected. Even in Dhaka, where one would expect a higher density, less than 50 percent of households adopted electricity, although 80 percent of villages had some type of electricity service avail- able (see Table 2.6). Summing Up This chapter has raised an important question: Are rural Bangladeshis moving up or down the energy ladder? In terms of household energy use, biomass in all its various forms is of paramount importance. For cooking, fuelwood is the predominant form of biomass used; Table 2.5. Energy Sources for Lighting (annual kgoe per household) Energy Type Chittagong Dhaka Khulna Rajshahi All Divisions Kerosene 27.86 18.64 25.30 17.62 22.38 Grid electricity 11.82 9.70 1.86 3.29 6.86 Candle 2.60 0.09 0.30 0.02 0.79 Solar PV 0 0.04 0.13 0 0.04 Storage cell 0 0.05 0 0 0.01 All 42.28 28.52 27.59 20.94 30.09 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 17 but it is becoming scarcer than in the past, Table 2.6. Electrification of Villages and Households by Division and a significant portion is being purchased. At the same time, many rural residents are Grid Connection (%) turning to residues, including grass and leaves, Division Villages Households to meet their daily cooking needs. Chittagong 78 39 In terms of lighting, kerosene is the main Dhaka 80 44 energy source, although most rural house- Khulna 50 11 holds would prefer electric lighting if it were Rajshahi 53 20 available. Electricity is not used as extensively All 66 29 as it might be, perhaps because of restrictive household connection policies. Source: BIDS Survey (2004). The energy-use situation of rural house- holds implies the need for policy initiatives that promote more efficient use of biomass energy, and its increased supply and more effi- cient use and better pricing of modern energy to increase supply and reliability. To gain a clearer picture of whether rural Bangladeshis are moving up or down the energy ladder, it is critical to examine patterns of energy demand (expenditure and asset ownership), which are the subject of the next chapter. Notes 1. Such fuels as unprocessed agricultural residue and cow dung are considered inferior to wood for cooking and generally produce higher levels of IAP when burned in traditional stoves. 2. The increase may involve advances in survey design instruments with which to measure rural energy use (e.g., previous surveys had more basic questions). CHAPTE R 3 Household Energy Demand I n Bangladesh, access to modern energy is a contentious issue, revolving around the role of energy markets. Although financially motivated rural shopkeepers should want to offer all types of commercial fuels, government policies determine, in part, which modern fuels are marketed in rural areas. A major question is how markets can function to serve rural people. To determine the extent to which modern energy factors in market development, this chapter examines patterns of rural household energy expenditure and asset ownership. Expenditure Patterns Household energy expenditure depends on many factors, from the availability and pricing of goods and services to household income. The following sections summarize the share of total rural energy expenditure by energy source and household income and region. By Energy Source As Figure 3.1 illustrates, rural residents still depend on traditional biomass, particularly fuelwood, to meet most of their energy needs. But the importance of commercial fuels far exceeds their basic energy content. Fuelwood accounts for 44 percent of total household energy use and 38 percent of cost. With regard to the commercial fuels purchased in rural areas, kerosene accounts for only 2 percent of total household energy use but 12 percent of cost, while electricity Note: In addition to the main authors, Hussain Samad was a contributing coauthor of this chapter. 19 20 World Bank Working Paper accounts for just 1 percent in terms of Figure 3.1. Energy Expenditure energy content, but 13 percent of cost Distribution by Energy Source (10 percent for grid-based electricity and (percent expenditure) 3 percent for solar PV, storage cells, and dry-cell batteries). The implied differences Non-grid in fuel and energy prices are discussed electricity 3% Other later in this chapter. 1% Grid electricity 10% By Income and Region In most cases, income level and fuel Kerosene Fuelwood expenditure exhibit a positive monotonic 12% 38% relationship (see Table 3.1). Even with regard to tree leaves and cow dung, where Dung the pattern is not completely monotonic, 14% higher-income households generally spend Tree leaves Crop residue 9% more than lower-income ones on energy. 13% Only for the "other" category, which is numerically unimportant, is there no clear Source: BIDS Survey (2004). pattern. Note: Costs are actual or imputed for each Poorer households spend less than energy source; consumer households must also bear the costs of procurement and supply. wealthier households on energy, but the Table 3.1. Annual Energy Expenditure by Income and Division (Tk per household) Income Category Tree Crop Cow Grid Non-grid (thousands of Tk) Fuelwood Leaves Residue Dung Kerosene Electricity Electricity Other <25 928 419 538 562 440 167 90 18 25­50 1,688 498 582 745 570 304 146 15 50­75 2,455 469 692 711 667 553 190 23 75­100 2,877 500 811 730 706 664 254 42 >100 3,016 454 779 875 822 1,288 293 194 Division Chittagong 3,763 300 352 479 732 761 185 124 Dhaka 1,611 713 823 654 560 710 136 10 Khulna 1,644 612 772 837 661 146 231 20 Rajshahi 685 279 648 922 474 284 140 14 All 1,962 470 641 716 608 488 172 44 Source: BIDS Survey (2004). Note: The annual energy expenditures include an imputed value for the collection time involved in gathering biomass fuels. The figures for expenditures on electricity and kerosene are reasonably close to those of the Bangladesh Household Income and Expenditure Survey (HIES). According to that 2005 national survey, the annual rural expenditure for kerosene was Tk 566 and Tk 522 for electricity. The results are close, given the different methods used to measure these variables. Restoring Balance: Bangladesh's Rural Energy Realities 21 amount represents a greater percentage of their income. The lowest-income households spend slightly more than Tk 3,000 or 15 percent of their annual income on energy. Higher-income households spend twice as much, but their income is more than four times higher. This pattern is common in other developing countries (World Bank 2002b). Cooking Energy: Biomass and Its Opportunity Cost Rural Bangladeshis not only spend a high proportion of their cash income on energy; they also collect substantial amounts of biomass from fields and local forests. The opportunity cost of biomass collection must be considered to understand the nature of this energy cost. In this context, regional differences are likely because of variations in amount of tree cover (e.g., Chittagong and Khulna have major forest lands). On average, households spend about 200 hours per year collecting biomass fuels (see Table 3.2). Valued at the average agricultural wage derived from the survey for both men and women, the annual value of this work equals about Tk 1,625 per family. When added to the energy expenses of the poorest households, expenditures increase about 50 percent. Collection time is evenly divided between gathering tree leaves and collecting fuelwood and cow dung. Crop residue involves little time overall, since it is generally collected as part of farm work. By contrast, residue collection is seasonal and thus periodically time intensive. For all types of biomass, the relationship between collection time and income is generally negative. One also observes substantial regional variations. For example, in forested areas of Chittagong, households spend much time collecting fuelwood; but in tree-deficient Rajshahi, more time is spent gathering tree leaves and cow dung. Women devote an average of 150 hours per year collecting biomass, disproportion- ately more time than men or children spend (see Figure 3.2).1 Children play a small role (probably collecting biomass from nearby the house). Previous studies appear to have Table 3.2. Biomass Collection Time by Income and Division (annual hours per household) Income Category Biomass Source (thousands of Tk) Fuelwood Cow Dung Tree Leaves Crop Residue <25 44.4 38.8 112.2 2.1 25­50 74.5 26.0 120.4 1.3 50­75 75.5 13.9 119.9 1.0 75­100 51.2 15.9 106.5 0.8 >100 31.9 19.7 98.1 0.7 Division Chittagong 152.9 20.7 94.9 1.0 Dhaka 28.4 14.9 116.1 0.8 Khulna 32.2 28.8 143.9 1.6 Rajshahi 20.4 34.6 106.3 1.8 All 60.6 24.6 114.3 1.3 Source: BIDS Survey (2004). 22 World Bank Working Paper Figure 3.2. Household Biomass Collection Time by Income Quintile (thousands of Tk per year) Source: BIDS Survey (2004). overestimated the burden of fuel collection by children, which findings from other recent studies on energy and time use confirm (World Bank 2002a). Lighting Energy Most rural Bangladeshis use kerosene as their primary lighting source. Some 70 percent of lighting energy is derived from kerosene, even though the quality of lighting service from electricity is an order of magnitude higher across all income groups and regions. In terms of pricing, kerosene varies little across income quintiles or regions. But for electricity, from which about 30 percent of lighting energy is derived, prices fall steadily as household incomes rise. Other factors that influence electricity pricing include connection type and regional variations. Electricity Connection and Consumption Electricity connection and consumption rates are influenced by a variety of factors, among which income level and regional differences (discussed next) figure prominently. A third factor (not discussed in this section) is proximity to the electricity supply line. Although one might expect that consumption of significant supplies may be beyond the reach of poor households, it may be that poor consumers living nearby the grid are being excluded from the potential development benefits of basic electricity services. Thus, the question is whether certain policies exclude poor households from connecting to the grid. Restoring Balance: Bangladesh's Rural Energy Realities 23 Table 3.3. Household Electrification Connection Rate and Consumption by Income Quintile Annual Income (thousands of Tk) Connection Rate (%) Household Consumption (kWh)* <25 15.0 32 25­50 23.2 76 50­75 33.7 167 75­100 38.2 180 >100 54.2 455 All quintiles 29.0 144 Source: BIDS Survey (2004). Note: The survey showed that the overall monthly electricity use for households with electricity is 41 kWh, which is close to the 38-kWh figure from the records of the Rural Electrification Board. *Household use = annual consumption per household for all survey respondents. Income Disparities. When other variables are held constant, the probability of having a household grid connection has been found to increase monotonically as income levels rise (Asaduzzaman and Latif 2005). For example, compared to the lowest income quintile (less than Tk 25,000), the next higher group (Tk 25,000­50,000) is 50 percent more likely to have an electricity connection (see Table 3.3). Thus, for poorer households, the high cost of obtaining a grid connection may be prohibitive. When connection and consumption disparities by income quintile are compared, one sees that disparity by income is higher for consumption because households without a connection are included in the consump- tion calculation. When this study considered the independent influence of income, results showed that the highest income quintile (more than Tk 100,000) was four times more likely than the lowest (less than Tk 25,000) to have electricity, and the likelihood increased with income. Regional Variations. Electricity connection and consumption rates also vary significantly by region. As Figure 3.3 illustrates, Dhaka and Chittagong have far higher percentages of households with electricity connections and, consequently, kilowatt-hours of consumption. Khulna exhibits the lowest percentage of connections and kilowatt-hours used. Kerosene versus Electricity Although 70 percent of lighting energy is derived from kerosene and only 30 percent from electricity, in terms of cost and kilolumen output, electricity is far superior to kerosene. Electricity's output per unit of energy is 100 times higher than that of kerosene.2 Trends. When one compares households that use kerosene lamps or electric lighting, two trends are obvious. First, level of lighting increases with income for both kerosene and electricity. Second, the quality of lighting service obtained from electricity is an order of magnitude higher than that from kerosene for all income groups across all regions. For the average household, the number of kilolumens is nearly 80 times higher for those who use mainly electricity, versus kerosene (see Table 3.4). The survey findings show that welfare gains from access to electric lighting are extremely high. Chapter 4 analyzes these benefits in more detail. 24 World Bank Working Paper Figure 3.3. Regional Variation in Electrification Source: BIDS Survey (2004). Price Differences. Prices paid for kerosene and electric lighting services differ markedly. For kerosene, prices for quantities used are fairly uniform, at Tk 20 to 21 per liter, with little variation across income quintiles or regions (see Table 3.5). When prices paid by level of lighting use are considered, households in the middle quintile (Tk 50,000­75,000) pay the lowest price per unit and poorest households (less than Tk 25,000) pay the highest price. But for electricity, prices fall steadily as household incomes rise. The price paid by the highest income quintile is less than half that paid by the lowest. By region, households Table 3.4. Comparison of Kerosene and Electric Lighting Services in Dhaka and Chittagong pay the least amount. Income Quintile Kerosene Lamps Electric Lights Because households with (thousands of Tk/year) (klm-hr/month) (klm-hr/month) electricity enjoy a much higher <25 2.99 130.30 quality of lighting than those that 25­50 4.00 218.22 use kerosene lamps, kerosene- 50­75 4.71 353.11 using households would need 75­100 4.97 329.73 to pay 74 times the amount paid >100 5.86 509.32 by households with electricity Division to attain the same illumination Chittagong 4.34 441.31 (see Table 3.5). This ratio rises Dhaka 3.44 297.67 as households move up the Khulna 4.46 209.57 income ladder; that is, higher- Rajshahi 4.81 202.40 income households pay a lower All 4.26 324.97 price for electric lighting than do poorer ones. Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 25 Table 3.5. Annual Price of Lighting by Income Quintile and Division Income Quintile Kerosene Kerosene Electricity Relative Price of (thousands of Tk) (Tk/liter) (Tk/klm-hr) (Tk/klm-hr) klm-hr (K:E) <25 21.2 46.1 0.71 64 25­50 21.1 39.1 0.54 72 50­75 20.9 27.7 0.43 64 75­100 20.9 35.5 0.40 88 >100 20.9 30.5 0.33 92 Division Chittagong 20.3 22.7 0.41 55 Dhaka 21.5 46.2 0.36 128 Khulna 20.9 43.0 0.61 70 Rajshahi 21.5 30.1 0.75 40 All 21.0 34.9 0.47 74 Source: BIDS Survey (2004). Price Variations of Electric Lighting In rural Bangladesh, about 75 percent of household connections are obtained directly from the PBSs; about 21 percent are from neighbors connected to the PBSs, while the remaining 4 percent are mainly from the Power Development Board (see Table A1.28, Annex 1). The average prices for kilowatt-hours and kilolumen-hours are therefore estimated by the type of connection and level of income (see Table 3.6). For PBS connections, the same monotonic and negative relationship with income is clearly discernible for price by kilowatt-hour and kilolumen-hour. The same is generally true for connections from neighbors. (Because the "other" category of connection sources has few observations, the results are probably not meaningful.) More interestingly, PBS-connected households pay the highest price for each income class, but the difference narrows sub- stantially as incomes rise. When divisional pricing is considered, Khulna and Rajshahi households are found to pay higher prices in nearly all cases. The increase in takas paid per kilowatt-hour is probably caused by a large first block for households in the PBSs, com- bined with fixed charges on electricity bills; as the amount of electricity increases, the price per kilowatt-hour decreases because fixed charges are averages across total kilowatt-hours used. Also, the marginal cost of a kilowatt-hour is less than the price charged to neighbors. But the profits taken by those who sell electricity to their neighbors appear negative.3 Such findings call for more in-depth analysis of formal and informal pricing of PBS connection arrangements. Asset Ownership Patterns Cooking, lighting, and non-lighting electric appliances all have important implications for rural well being. Biomass cookstoves, for example, are closely tied to levels of indoor air pollution (IAP) and negative effects on human health. The superior lighting quality of electric 26 World Bank Working Paper Table 3.6. Price of Electric Lighting by Income, Connection Type, and Division Electricity Price Electricity Price Income Quintile (Tk/kWh) (Tk/klm-hr) Average Price (thousands of Tk) PBS PBS-N Other PBS PBS-N Other Tk/kWh Tk/klm-hr <25 9.2 5.8 4.1 0.87 0.48 0.34 7.7 0.71 25­50 7.2 4.7 2.9 0.63 0.40 0.23 6.3 0.54 50­75 5.4 4.1 4.6 0.47 0.36 0.28 5.0 0.43 75­100 5.1 5.4 1.7 0.40 0.46 0.14 5.1 0.40 >100 4.3 3.4 3.2 0.34 0.28 0.23 4.2 0.33 Division Chittagong 5.4 4.4 3.8 0.44 0.36 0.26 5.1 0.41 Dhaka 4.7 4.5 2.8 0.35 0.38 0.21 4.6 0.36 Khulna 7.2 4.3 3.5 0.69 0.41 0.28 6.4 0.61 Rajshahi 8.1 5.4 -- 0.79 0.50 -- 7.7 0.75 All 5.9 4.6 3.7 0.50 0.39 0.26 5.5 0.47 Source: BIDS Survey (2004). Note: PBS-N = indirect connection to the PBS via neighbor (N). The PBS-Ns are estimates based on appliance ownership, as no bills are available and electricity is often based on fixed rates. lamps is closely tied to evening study and educational improvements. Electric fans provide indoor space cooling, while radios and televisions offer communication and entertainment. Thus, it is important to understand typical rural household ownership patterns of such appliances. Cookstoves and Health Implications Across all regional divisions, fixed clay cookstoves are used for burning all types of biomass.4 In the more highly developed division of Dhaka, nearly 7 percent of households use kerosene stoves. On average, households own two clay stoves (see Table 3.7). In Dhaka and Rajshahi, Table 3.7. Household Ownership of Cooking Stoves by Type (number per 100 households) Division Stove Type Chittagong Dhaka Khulna Rajshahi Total Clay (fixed) 182.8 200.0 174.8 172.3 182.7 Clay (portable) 3.9 33.2 12.6 32.8 20.5 Kerosene 0.8 7.1 1.3 1.0 2.5 Gas 1.3 0.1 -- -- 0.4 Electric heater 1.7 0.3 -- 0.2 0.2 Total 190.5 240.7 188.7 206.3 206.3 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 27 about 33 percent of households own portable Table 3.8. Average Effective Cooking Hours by Income Quintile stoves, which may be used for cooking out- doors. A few households also use kerosene Income Quintile stoves, while gas and electric stoves are rare. (thousands of Tk) Total Cooking Hours Average cooking time is three to four hours <25 4.2 per day using fixed stoves and about two 25­50 5.9 hours a day using portable ones. 50­75 7.0 The average clay stove, whether fixed or 75­100 8.0 portable, is used two to three hours daily. >100 9.2 However, when total effective hours (number Average 6.4 of stoves times number of hours of operation) Source: BIDS Survey (2004). are considered, wealthier households are found to cook for longer periods of time than poorer ones (see Table 3.8). As discussed in Chapter 4, IAP is a major health risk for women and children. It may present an even more serious hazard for wealthier households, given their longer cooking hours. At the same time, poorer households, who cook only once or twice a day, often end up eating cold meals more frequently than wealthier households, who cook two to three meals per day. In short, the net effect of IAP on human health and nutrition is difficult to predict. Lighting Appliances Kerosene Lamps. Bangladeshi households use two major types of kerosene appliances: kupis and hurricane lamps. Typically, kupis are uncovered lamps with a single handmade wick made of discarded fabric. Hurricane lamps, which have a glass chimney and thicker (often purchased) wicks, provide brighter light than kupis. In addition to these types, pressurized kerosene lamps, called petromaxes or hachaks, provide even brighter light. Kupi ownership is virtually universal, while hurricane lamps are owned by far fewer households (see Table 3.9). Poor residents tend to rely more on kupis, which emit little light. The proportion of households who own hurricane lamps is similar across divisions, probably indicating their function as backup lighting for households with electricity. Electric Lamps. Virtually all households with electricity have electric lamps (using 60-W incandescent bulbs), fluorescent tubes, and compact fluorescent lamps. The number of households using more efficient fluorescent lamps is less than those using incandescent bulbs. In fact, ownership of fluorescent lamps is limited almost exclusively to the two highest income quintiles (see Table 3.10). Rural Bangladesh has virtually no compact fluorescent lamps. These patterns are understandable, given the large price differentials between an incandescent bulb and fluorescent tube or compact fluorescent lamps using similar wattage. Ownership of electric lamps thus depends on both the rate of rural electrification and household income. The number of lighting watts per household also increases by income level. The wealthiest quintile has electric lights with a total of 156 watts per household, nearly 10 times more than the poorest quintile. In divisions with low rates of electricity use, the average number of watts per household is typically low, generally reflecting the respective region's rate of rural electrification. 28 World Bank Working Paper Table 3.9. Household Ownership of Lighting Appliances by Electrification Status (number per 100 households) Household Appliance Type Status by Hurricane/ Charger Division Kupi/Cherag Lantern Petromax Light Bulb Tube Light (with torch) Households with electricity Chittagong 212.4 95.6 -- 414.5 77.9 19.3 Dhaka 153.0 80.3 1.1 310.6 36.7 -- Khulna 161.0 81.4 -- 308.5 20.3 1.7 Rajshahi 162.3 77.0 1.6 315.6 18.9 0.8 All 176.7 85.3 0.7 348.6 47.0 7.2 Households without electricity Chittagong 210.2 98.2 -- -- -- 0.5 Dhaka 162.5 89.7 0.6 -- 2.9 -- Khulna 196.7 84.7 0.4 -- 6.7 -- Rajshahi 160.5 90.2 0.8 -- -- -- All 182.8 90.3 0.5 -- 2.5 0.1 All households Chittagong 211.1 97.2 -- 161.3 30.3 7.8 Dhaka 158.4 85.6 0.8 136.0 17.7 -- Khulna 192.9 84.3 0.4 33.2 8.2 0.2 Rajshahi 160.8 87.5 1.0 64.2 3.8 0.2 All 181.0 88.9 0.5 101.2 15.4 2.2 Source: BIDS Survey (2004). Non-lighting Electric Appliances Given Bangladesh's hot tropical climate, it is not surprising that the most frequently observed non-lighting appliance in households with electricity is the fan (see Table 3.11). Two out of three households own at least one electric fan. One out of three households owns a television set. Ownership of fans, radios, and television sets is highly correlated with income level. With few exceptions, fans are found only in households with electric- ity. Some households without electricity own and operate battery-powered radios and television sets. Demand for Other Energy: Quantitative Analysis The cross-sectional analysis from the previous sections suggests that factors such as income and landholding determine the patterns and types of energy used by rural households. Additional variables that may influence household energy demand include community infrastructure and consumer prices. Together, such factors affect energy demand for both household consumption and production. For example, both income and prices influence Table 3.10. Electric Lighting by Income and Division Income No. Compact Quintile Households Total % Using Compact Fluorescent (thousands with Electric Wattage/ % Using No. Bulbs/ Bulb Fluorescent Fluorescent Lamp of Tk) Lighting (%) Household Bulbs Household Wattage Lamps Lamps Wattage <25 15.0 16 14.8 0.3 16 0.6 0.01 0.5 25­50 23.2 35 23.1 0.6 33 2.7 0.04 1.4 Restoring Balance: Bangladesh's Rural Energy Realities 50­75 33.7 69 33.7 1.1 62 9.2 0.16 6.1 75­100 38.2 79 38.7 1.4 69 12.4 0.31 10.0 >100 54.2 156 53.9 2.7 137 24.3 0.55 20.0 Division Chittagong 38.9 94 38.8 1.6 84 13.8 0.29 10.6 Dhaka 43.8 86 43.4 1.4 79 9.6 0.17 6.4 Khulna 10.8 17 10.9 0.3 15 2.9 0.08 2.0 Rajshahi 20.3 27 20.3 0.6 25 2.0 0.04 1.4 Average 29.0 58 28.9 1.0 52 7.3 0.15 5.2 Source: BIDS Survey (2004). 29 30 World Bank Working Paper Table 3.11. Ownership of Non-lighting Electric Appliances by Income Quintile Income Electric Fans Television Sets Radios/Tape Recorders Quintile Mean Mean Mean (thousands % % Number % Number % Number of Tk) Connected Ownership Owned Ownership Owned Ownership Owned <25 15.0 40.0 0.53 16.0 0.16 17.3 0.17 25­50 23.2 52.4 0.75 23.6 0.24 26.0 0.26 50­75 33.7 61.0 1.07 36.4 0.36 37.0 0.38 75­100 38.2 71.1 1.42 49.4 0.49 43.4 0.45 >100 54.2 73.0 1.87 59.2 0.62 50.0 0.55 All quintiles 29.0 60.4 1.16 37.6 0.38 35.6 0.37 Source: BIDS Survey (2004). farmers' selection of fuels and irrigation pumps. From a policy perspective, it is critical to determine the relative importance of these factors, given the competing demand for alternative sources of energy and its quality. The question for researchers is this: What are the direction and magnitude of these factors' effects on demand for various types of energy? Table 3.12. Summary Statistics of Outcomes To answer this question, we con- and Important Explanatory ducted an econometric analysis to pre- Variables of Household Energy dict how household energy demand is Demand Regressions affected by changes in the various factors. Standard A tobit regression was run to account Variable Type Mean Deviation for households having zero values for Household energy demand one or more types of energy demand. Fuelwood (kg/month) 98.86 104.57 The variables used as influencing factors Kerosene (liter/month) 2.41 1.94 were gender of household head, age, Diesel (liter/month) 1.80 11.18 maximum level of education in the Electricity (kWh/month) 25.72 252.29 household, household income (proxied Explanatory by land and non-land assets), com- Maximum education of 5.02 4.17 munity prices for major energy sources household adults (years) and consumer goods, and community Household assets infrastructure variables.5 Table 3.12 Land (acres) 1.25 2.22 provides summary statistics of energy Non-land (Tk 10,000) 1.68 5.55 demand and selected influencing vari- Village price ables of policy relevance, while Table 3.13 Fuelwood (Tk/kg) 1.65 0.68 presents the effects of selected variables Kerosene (Tk/liter) 22.94 2.73 on energy demand. Diesel (Tk/liter) 23.19 2.61 The study results support the Electricity (Tk/kWh) 2.85 0.40 assumption that an increase in income If village has electricity 0.66 0.47 increases the demand for energy. Indeed, with higher incomes, rural Bangladeshis Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 31 Table 3.13. Estimates of Household Energy Demand Household Demand (monthly) Fuelwood Kerosene Diesel Electricity Explanatory Variable (kg/month) (liter/month) (liter/month) (kWh/month) Household Maximum education of 1.46** 0.020* 0.001 1.49 adult males (years) Maximum education of 0.76 0.020 -0.001 2.38 adult females (years) Land assets (acres) 1.91** 0.080** 0.020** 3.71 Non-land assets (Tk 10,000) 1.55** 0.020** 0.020** 2.65** Village price Fuelwood (Tk/kg) -34.73** 0.320** -0.003 -0.12 Kerosene (Tk/liter) 5.95** -0.009 0.002 0.53 Diesel (Tk/liter) -8.81** -0.110** 0.001 -0.41 If village has electricity 14.21** -0.600** -0.020 32.78** Source: BIDS Survey (2004). * = significance level of 10 percent; ** = significance level of 5 percent or stronger. Figures represent changes in energy demand for fuelwood, kerosene, diesel, and electricity caused by unit changes in the explanatory variables. were found to pursue improved sources of energy and invest in assets that require more energy. Household income, apportioned by landholding and non-land assets, influences the demand for fuelwood and diesel, but non-land assets matter most in demand for elec- tricity. It should be noted that income flow can be influenced by electricity consumption, where electricity use is viewed as an input in production. By contrast, assets are stocks that do not change over the short term but only over the long run. Therefore, we use assets rather than income to estimate the effect of household income proxies on energy use. For example, an increase of 1 acre in landholding increases household consumption of fuel- wood by 1.91 kg per month and kerosene consumption by 0.08 liters per month. An addi- tional Tk 10,000 of non-land assets increases household electricity consumption 2.65 kWh per month. As incomes increase, rural households typically diversify their energy portfolios, moving toward modern energy, which is not only more efficient but also allows household members to invest time and money in more productive activities than biomass energy alone would make possible. The study findings confirm that education reflects rural people's preference for improved sources of energy. For example, an additional year of education for household adult males increases electricity consumption by nearly 1.46 kWh per month. Chapter 4 discusses in more detail the benefits of the transition to modern energy services (e.g., from biomass to kerosene or kerosene to electricity) on overall household welfare and productivity. The price of an energy source has a direct effect on its own consumption and a cross-effect on the demand for alternative energy sources, depending on whether they are substitutes. For example, a price increase in fuelwood leads to decreased demand for it (direct effect) 32 World Bank Working Paper and increased demand for kerosene (positive cross-effect). A similar positive cross-effect is observed with regard to fuelwood demand when the price of kerosene rises. These phenom- ena clearly suggest that, for rural households, fuelwood and kerosene are substitutes. The presence of electricity also increases fuelwood demand. That fuelwood is considered a good fuel for cooking in Bangladesh supports the notion that villages with electricity are more affluent than those without electricity. These measures have an overall village-level income effect on the demand for alternative energy sources. In addition, availability of electricity results in reducing villagewide energy prices, which affords households more to spend on energy generally, including fuelwood. But demand for kerosene declines with village electrification, which signals the substitution effect of users switching from kerosene to electricity (mostly for lighting). Summing Up In rural Bangladesh, access to energy is governed not only by its availability, but also by its pricing (monetary and non-monetary), household income, and other characteristics. Household demand for energy is sensitive to both price and income. If electricity is available, household access is determined by income; the higher the income (proxied by land and non-land assets), the greater the demand. For any energy source, pricing also determines demand. Thus, energy access is not equitable because the source is either priced too high or carries a high opportunity cost (e.g., biomass collection). Inequitable access and level of use may lead to further inequities (e.g., low-income households may pay far higher prices for equivalent lighting services). By exploiting price sensitivity, policy makers can play a role in influencing demand (e.g., by withdrawing the price subsidy for an energy source whose demand is insensitive to price or, conversely, introducing a price subsidy for a source whose demand is negatively sensitive to price). The next chapter demonstrates the loss to household welfare that can result from such inequities and the important gains that moving up the energy ladder makes possible. Notes 1. Much of women's time is spent collecting tree leaves and grass; in terms of fuelwood collection, men's time is at least as important as that of women. 2. Efficiency of appliances (kerosene kupis and hurricane lamps or electric bulbs and tubes) deter- mines the ultimate lighting use rate, which, in turn, determines the quantity of fuel used. 3. Households that sell electricity to their neighbors cannot accurately determine the amount of such use. The neighbor buyers, who do not have meters, usually pay a fixed monthly amount based on a rough estimate of load, which they underestimate or under-report to the household sellers. The actual load is determined by the wattage and hours of use of all electric appliances and lightbulbs, which is difficult for the sellers to compute. 4. In such locations as Sylhet District, sun-drying, rather than parboiling, is used to process paddy before milling. 5. Only those influencing variables considered most important are reported here; see Table A2.1 (Annex 2) for the complete regression output. CHAPTE R 4 Household Gains from Energy Use A cross the developing world, rural residents have benefited from switching to electricity and other forms of modern energy. Productivity gains in agriculture and business have increased household welfare directly, while skills development and improved education have contributed indirectly to a higher quality of life. This chapter first considers the gains in Bangladeshi household welfare and farm pro- ductivity that result from moving up the energy ladder. Given Bangladesh's heavy reliance on biomass cooking energy, the health risk of indoor air pollution (IAP) is analyzed, and potential mitigation measures are considered. Next, the direct and indirect benefits of energy use on household income and consumption are presented within the context of the energy ladder typology. Biomass Cooking: From Traditional to Modern Energy Some 95 percent of Bangladeshi households collect or purchase biomass energy to cook all or part of their meals, mainly using fixed clay stoves. Although biomass is freely available or inexpensive, the inherent inefficiency of stoves, combined with the high moisture content of biomass cooking fuels, results in incomplete combustion, which produces excessive smoke. Indoor Air Pollution from Traditional Stoves: Empirical Evidence Long-term exposure to smoke and small particulates has long been considered an environ- mental health risk. Although there are many ways to measure exposure to smoke, the most In addition to the main authors, Hussain Samad and Susmita Dasgupta were contributing coauthors of this chapter. 33 34 World Bank Working Paper common method involves the measurement of small particulates. Small particulates are a heath concern because they have an impact deep in the lungs and are a major cause of respiratory illness. The most general measurements of particulate matter are called PM, with a number afterwards that represents the aerodynamic diameter of the particle. For example, PM10 refers to fine particulate matter with an aerodynamic diameter smaller than 10 micrograms (µg). PM2.5 has an aerodynamic diameter smaller than 2.5 µg. The exposure levels recommended by health and environmental agencies generally refer to the number of particulates of a certain diameter per cubic meter of air. The most recent standards of the World Health Organization for PM10 are 20 µg/m3 of air annual mean exposure levels and 50 µg/m3 24-hour mean exposure levels (WHO 2006). For PM2.5, the recommended levels are 10 µg/m3 annual mean exposure levels and 25 µg/m3 24-hour mean exposure levels. For Bangladesh, the standard for annual recommended average for PM10 is 50 µg/m3. Worldwide evidence indicates that women and their children bear the brunt of indoor air pollution (IAP). Moreover, poorer households, compared to wealthier ones, bear the heavier burden. In rural Bangladesh, age­sex composition of households indicates that generally women and children suffer greatest exposure. Adult men are the least exposed because they spend more time outside the home. Solid particulates, especially those with diameters smaller than 10 microns (PM10) are released into the air; and, if inhaled for prolonged periods, can lead to various diseases, loss of health, and early morbidity. Finer particles (PM2.5) are even more damaging. This section discusses various aspects of IAP caused by incomplete biomass combustion. Both male and female children spend many hours indoors in the cooking space or adjacent living area until 6 to 8 years of age, when gender-based patterns begin to diverge (see Table 4.1). Adolescent males begin to spend more time outdoors, while adolescent females move in the opposite direction, spending more time in indoor cooking and living areas. Women 20 to 60 years of age spend nearly 4 hours per day cooking, nearly 16 hours in the household living area, and the remaining hours outdoors. For those above 60 years of age, the patterns are somewhat reversed. Table 4.1. Mean Daily Hours in Household Location by Age Group* Female Male Female­Male Age Cooking Living Cooking Living Cooking Living Group Area Area Outdoors Area Area Outdoors Area Area Outdoors 0­1 1.12 20.04 3.01 1.11 19.29 3.57 -0.01 -0.75 0.56 2­5 1.08 18.44 4.52 0.93 18.13 4.97 -0.15 -0.31 0.45 6­8 1.01 16.40 6.61 0.48 16.41 7.17 -0.52 0.01 0.56 9­12 1.32 15.55 7.19 0.31 15.61 8.06 -1.01 0.06 0.87 13­19 2.38 15.71 5.97 0.28 14.33 9.41 -2.10 -1.38 3.44 20­60 3.75 16.05 4.27 0.19 13.07 10.79 -3.56 -2.98 6.52 >60 1.48 19.76 2.93 0.17 16.56 7.37 -1.31 -3.20 4.45 Source: Dasgupta et al. (2004). *For seven regions in Bangladesh. Restoring Balance: Bangladesh's Rural Energy Realities 35 Evidence suggests that pollution levels in living areas adjacent to the kitchen are broadly similar. Except for a few hours in the early morning and early afternoon, PM10 concentrations in kitchen and living areas remain far above 100 µg/m3, the Indian standard for air pollution load. Ambient concentration, for the most part, remains below the Indian standard, except during early evening hours (see Figure 4.1). As Dasgupta et al. (2004) have estimated, both males and females are exposed to solid particulate pollution far above accepted standards for all ages, except for males 20 to 60 years of age, whose exposure level is little above the standard pollution concentration (see Figure 4.2); but they, too, are at risk for respiratory and other diseases. Exposure levels and their consequences may be tempered by various socioeconomic factors, including education and income, which influence awareness, fuel choice, mitigation measures, and general living environment. Among Bangladeshi children, acute respiratory infection (ARI) remains a major cause of morbidity and mortality. A 2004 survey reported that 21 percent of rural children 1 to 5 years of age suffered from ARI.1 This study's household survey indicates a broadly similar picture. However, the survey data are imprecise. Because they only reported whether working days were lost in the prior month due to specific diseases, the reported incidence of children's diseases may have been lower. In addition, the specific symptoms or diseases reported by household heads were subject to bias. Because symptoms were not always well clarified, it is difficult to know whether respiratory problems were present.2 Despite these information gaps, the findings are instructive. Overall, 10 to 11 percent of men and women experienced respiratory problems and diseases. About 24 percent of male children and 20 percent of female children in the 0 to 5 age group suffered from such Figure 4.1. PM10 Concentration by Time of Day and Household Location Source: Dasgupta et al. (2004). 36 World Bank Working Paper Figure 4.2. PM10 Concentration by Age Group and Gender Source: Dasgupta et al. (2004). diseases. For children and adolescents 10 to 19 years of age, about 8 percent of both genders were affected. For adults 20 to 60 years of age, the proportions were 9 percent for men and 10 percent for women. Thus, there was little difference by sex, but young children suffered more than adolescents and adults. Evidence from rural Bangladesh indicates that IAP varies by fuel type, cooking location, kitchen ventilation characteristics, and other factors; furthermore, non-fuel factors may be as important as fuel characteristics (Dasgupta et al. 2004). The extent and duration of particulates in the kitchen, as well as the amount of smoke that leaks into other living spaces or the outdoors, may depend on kitchen location and ventilation, as well as the porous quality of materials used to construct kitchen roofing and walls. This survey revealed wide variation in household-specific ventilation characteristics. Regression of 24-hour PM10 concentration on fuel use and a large set of variables that describe household cooking and ventilation practices suggest that ventilation factors play a larger role than fuel choice. Two construction materials, mud walls and thatch roofs, significantly affect PM10 concentration.3 If the cooking location is indoors, the sealing effect of mud walls increases PM10 concentration significantly. After controlling for stove locations, construction materials, and opening of doors and windows after cooking, the PM10 effects of biomass fuels (dung, fuelwood, twigs and tree branches, dung, rice husks, and straw) are indistinguishable. Relative to these fuels, which are taken as the baseline, use of kerosene subtracts about 90 µg/m3, LPG or LNG subtracts 103 µg/m3, and piped natural gas 136 µg/m3 from indoor PM10 concentration. Figure 4.3 presents the interactive effect of critical pollution factors by computing mean PM10 con- centrations for groups that distinguish between clean (kerosene and natural gas) and biomass fuels, inside and outside (detached or open-air) cooking, and mud-wall and other construc- tion materials. For statistical comparison of means, the benchmark is the mean living-space PM10 concentration for households with biomass fuels, inside cooking, and non­mudwall construction (acronym BIOL). The mean concentration for these households is 223 µg/m3. Restoring Balance: Bangladesh's Rural Energy Realities 37 Figure 4.3. Pollution Factors and PM10 Concentrations Source: Dasgupta et al. (2004). Pitt, Rosenzweig, and Hassan (2005), working with Bangladesh data, show that such exposures may not be random. Their findings show that cooking is done mostly by women. But those women in poorer health are more involved in cooking, possibly because they have little chance to switch to cleaner fuels or because the task falls onto those who otherwise cannot contribute to more productive activities. Among mothers, those with young children are less involved in cooking than others, indicating that cooks exposed to IAP are conscious of its hazards and avoid them as much as possible. Yet, because fuel-substitution possibilities are extremely limited or costly, the adjustments are insufficient. This study has detected dangerously high levels of pollution in rural Bangladeshi households. Concentrations of respirable airborne particulates (PM10) 300 µg/m3 or greater are common, implying widespread exposure to a serious health hazard. Within households, individuals' exposure is related to concentrations of pollution in indoor locations during the daily round of activities. The estimates reveal high levels of exposure for children and adoles- cents of both sexes, with children under five years of age at increased risk. Among prime-aged adults, men have half the exposure of women (whose exposure is similar to that of children and adolescents). Elderly men also have significantly lower exposure than elderly women. 38 World Bank Working Paper Predictably, fuel choice affects indoor pollution levels. Kerosene and natural gas are significantly cleaner than biomass fuels. But econometric results strongly suggest that certain household characteristics--construction materials, space configurations, cooking locations, and use of doors and windows--matter as much or perhaps more than fuel choice in deter- mining PM10 concentrations. In some biomass-burning households, concentrations are scarcely higher than in those using natural gas. Across households, family income and adult education levels (particularly that of women) significantly affect choices of cooking fuels, cooking locations, construction materials, and ventilation practices. As a result, the poorest, least-educated households have twice the pollution levels of higher-income households with highly educated adults. The exposure of young children and poorly educated women in poor households is quadruple that of men in higher-income households organized by more highly educated women. Possible Gains from Cooking with Modern Stoves Use of improved stoves and more modern forms of cooking energy has several classes of benefits, some of which are private and others public. This section considers mainly the private benefits, which are fairly self-evident. Because improved stoves require less fuel, using them reduces either the time spent collecting biomass fuels or the money spent to purchase them. Cooking with modern fuels involves cash expenses, but also entails many benefits. In most rural economies, biomass fuels are collected from the local environment; thus, the benefits mainly involve the reduction in collection time for household members, especially women. As reported in Chapter 3, annual biomass-collection time for rural households in Bangladesh totals about 200 hours (150 of which are contributed by women). This means that using an improved stove saves about 25 percent of both fuel and collection time, representing a significant annual savings. For those that purchase biomass fuels, the benefits are similar to, but more straightforward than, time savings. If 25 percent less fuel is used for the same amount of cooking, then avoided expenses equals a 25 percent monetary gain. Similar potential gains can result from using modern fuels, such as LPG, for cooking. It is not as straightforward to estimate the private health benefits of using improved stoves and modern cooking fuels (e.g., avoided eye irritation and illnesses or early deaths caused by exposure to high levels of IAP and improved cleanliness of the home). Most households are well aware of the benefits of eliminating smoke from their kitchens and other rooms. Household members, especially women, know the drudgery involved in keeping pans and walls free of the extensive carbon produced by traditional stoves. Assuming the same level of service and ease of use, households should, at least theoretically, be willing to pay more for a stove that involves less eye and throat irritation and less cleaning of pots and pans. However, the health and avoided death benefits are more difficult to grasp.4 Despite these difficulties, the World Health Organization is making efforts to quantify some of the public and private benefits of using improved stoves or substituting petroleum for biomass cooking fuels (Hutton and Rehfuess 2006; Hutton et al. 2006; WHO 2006). These studies develop economic valuation methods to evaluate time savings involving fuel collection and cooking, avoided health costs, and environmental benefits, among others. For a typical household in South Asia, the benefits of switching exclusively to improved stoves or from biomass energy to LPG or kerosene total about US$30 per year. Some of the Restoring Balance: Bangladesh's Rural Energy Realities 39 best new-generation models, lasting 5 to 10 years, cost about US$50. Thus, households that use such stoves can expect a positive return on investment within two years. The benefits of improved stoves, with certain caveats, significantly outweigh the costs, at least in theory. Unfortunately, the rural reality tells a much different story. If rural households perceived that even a fraction of the benefits developed under the national program were true, the program would not require significant subsides. Rural residents would be willing to purchase such stoves at their retail cost. But improved stoves must be designed to deliver a range of cooking services that people want: fuel efficiency, reduction in IAP, ability to cook a wide array of food preparations, cooking convenience, and even attractiveness in the kitchen or room. Given the history of low adoption rates, it is clear that past government subsidies used to promote the program have been problematic. Despite the effectiveness of improved stoves to reduce indoor cooking smoke, programs from around the world reveal a wide range of successes and failures. Obviously, if a household decides to revert to a traditional stove or uses an improved stove that develops cracks and leaks, the desired improvements in cooking efficiency and reduced IAP will not be achieved. In addition, even after a household adopts an improved stove, indoor pollution levels remain significantly higher than those advocated by international development organizations, such as the World Health Organization and environmental agencies. Clearly, Bangladesh requires more options with regard to the types of improved stoves promoted and potential use of modern cooking fuels. From Kerosene to Electric Lighting The high-quality lighting afforded by electricity has important consequences for Bangladeshi household welfare. This study applied the concept of consumer's surplus to measure the welfare gains for lighting resulting from transitioning from less efficient kerosene-using devices to those using electricity to obtain the same amount of benefit (Annex 3). Similar measurements could be made for other services (e.g., entertainment), which were beyond the scope of this study. Gains in Education Mounting global evidence suggests that rural electrification is closely tied to higher income and improved education. Various studies have confirmed the link between rural electrifi- cation and education, suggesting that children's school attendance improves as households adopt electricity (Kulkarni, Barnes, and Parodi 2007). Many rural surveys have found that children study better when the high-quality lighting made possible by electricity is available (World Bank 2002b). A recent study on rural electrification and development in Bangladesh indicates that households with electricity have higher literacy rates and increased school enrollment than those without electricity (Barkat et al. 2002). Evidence from this study suggests that the number of school-going students in rural Bangladeshi households influences the amount of electricity consumed (Asaduzzaman and Latif 2005). If the number of students in a family household increases by one, annual electricity consumption grows by about 51 kWh. This change is large, given that average annual household consumption is 114 kWh for domestic uses and 81 kWh for lighting. 40 World Bank Working Paper Table 4.2. Annual Energy Use by Number of Students in Household Lighting Domestic Uses Students in Kerosene Grid Electricity Kerosene Grid Electricity Household (no.) (liter) (kWh) (liter) (kWh) 0 23 52 25 92 1­2 28 84 30 153 3­4 34 121 36 209 >4 37 232 40 415 Average 27 81 29 144 Source: BIDS Survey (2004). Indeed, compared to electric lighting, use of kerosene lamps varies considerably less by number of students (see Table 4.2). The most plausible explanation for such increases in electricity use is that school- children use electric lights to extend their study time. According to the survey, students in households with electricity study longer (see Table 4.3). Total hours of study time (both daily and evening-hour) are greater in households that use electric lighting versus those without electricity, which use kerosene lamps. Clearly, electricity contributes to human-capital formation, leading to a longer-term increase in productivity. Moving to Better Lighting: Consumer's Surplus In rural Bangladesh, the most common household lighting appliances are kerosene-using kupi and hurricane lamps and electricity-using incandescent bulbs, fluorescent tubes, and compact fluorescent lamps. In this study, households were categorized into three groups: (1) those that light only with kupi lamps, (2) those with hurricane lamps that may or may not own a kupi lamp but do not have electricity, and (3) those with electricity that sometimes use an electric lamp and may or may not use any type of kerosene lamp. The gains in con- sumer's surplus were estimated for the transition from the kupi- to the hurricane-based system, and from the hurricane- to the electricity-based system. To understand the results of the estimates of welfare gains, it is necessary to observe the distribution of households by their lamp-holding characteristics. This study's findings indi- cate that kupi-based households are concentrated more among poorer households (see Table 4.4). Table 4.3. Study Time by Household Hurricane lamp holdings are far more common Electrification Status than kupis, and patterns across income classes are (average number of hours) not clearly discernible. Households with electric Household Status lamps are about 29 percent of households, Student With Without reflecting those that have electricity; and their Study Time Electricity Electricity proportion increases with income. Daily 2.72 2.13 One conservative measure of the improve- Evening 1.32 0.96 ment in consumer welfare is avoided expendi- tures from switching from kerosene to electricity Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 41 Table 4.4. Households by Lighting System and Income Income Quintile Type of Home Lighting System (% households) (thousands of Tk/year) Kupi Only Hurricane-based Electricity-based < 25 41.4 43.6 15.1 25­50 19.5 57.3 23.2 50­75 13.3 53.0 33.7 75­100 10.6 51.2 38.2 >100 5.0 40.6 54.4 Average 20.1 50.8 29.0 Source: BIDS Survey (2004). for lighting. Although this is a rather rudimentary measure, it is interesting to compare kerosene expenditures for households with and without electricity. Because households without electricity spend more on kerosene, some fuel substitution occurs (see Table 4.5). Generally, households with electricity spend 200 to 500 takas less per year on kerosene. Although this avoided expenditure is an interesting benefit, it severely underestimates the true benefits because households with electricity have significantly more light compared to those with kerosene lamps. A better way to measure the welfare gains of switching from kerosene to electric- ity involves a concept called consumer's surplus, which is based on a demand survey for lighting (see Annex 3). The gain in welfare, as measured by consumer's surplus, is substantial for the average consumer switching from any of the kerosene lamps to electric lights (see Table 4.6). The estimated benefit is 40 to 45 percent of household income (a substantial gain in terms of Bangladeshi takas). The reason is the inherent efficiency of electric lamps. The price of each kilolumen-hour from incandescent bulbs or fluorescent tubes is only about 1.43 percent of the cost of comparable lighting service from kerosene lamps (see Table 3.5). Table 4.5. Household Expenditure by Electrification Status on Kerosene and Other Energy Sources, by Income Group (Tk per year) Households without Electricity Households with Electricity Income Category Kerosene Other Energy Kerosene Other Energy (thousands of Tk) Expenditure Expenditure Expenditure Expenditure <25 467.8 2,466.0 281.9 3,054.7 25­50 645.9 3,613.8 319.2 3,865.4 50­75 794.4 4,715.2 415.6 4,179.2 75­100 883.6 5,886.1 418.8 4,095.9 >100 1,133.8 5,944.9 559.9 5,270.3 All 688.9 3,905.0 408.8 4,227.2 Source: BIDS Survey (2004). 42 World Bank Working Paper Table 4.6. Welfare Gains from Switching Lighting Systems System Transition Income Group Kupi Only to Hurricane- to Kupi Only to (thousands of Tk/year) Hurricane-based Electricity-based Electricity-based <25 56 (4.0) 838 (59.4) 894 (63.5) 25­50 67 (2.2) 1,333 (43.4) 1,401 (45.6) 50­75 83 (1.7) 2,243 (44.0) 2,326 (45.6) 75­100 58 (0.8) 2,156 (29.8) 2,214 (30.6) >100 80 (0.5) 3,197 (20.9) 3,276 (21.4) Average 72 (1.4) 2,057 (40.1) 2,129 (41.5) Source: BIDS Survey (2004). Note: Left-hand figures represent takas per month; right-hand figures in parentheses are ratios of the gains in consumer's surplus to the relevant group's average monthly income. Interestingly, when consumers are divided into income classes, the relative income gains are higher for poorer groups. But as expected, because of the higher density of lamps found in wealthier households, their absolute gains are two to three times higher than those of poorer groups. These findings have major rural energy policy implications for Bangladesh. From Manual to Mechanized Farming This study hypothesized that using modern energy for farming ultimately leads to gains in agricultural productivity. According to the United Nations Millennium Project: "At the household scale, modern energy services directly contribute to economic growth and poverty reduction. They create opportunities for income generation, reduce unit costs, and enable increased income from agriculture or animal husbandry by permitting pumping for supplementary irrigation, which lessens the risks associated with rainfed systems and enables increasing crop and pasture productivity, as well as switching to higher-value crops" (United Nations 2005). One should interpret the gains discussed here with caution, however, as more intensive data is needed to conduct an appropriate analysis.5, 6 Improved Productivity Bangladesh's agriculture is characterized by three rice-cropping seasons (Chapter 5). From the viewpoint of energy inputs, the most important is the dry season, when irrigated winter rice (boro) is grown; boro rice accounts for roughly 50 percent of rice production. The other is the partly wet/partly dry season, when aman rice, which may require supple- mentary irrigation, is grown. In all cropping seasons, both bullocks and power tillers are used for land preparation. As a result, four basic types of energy-using technology are classified by energy-use importance in production: low tillage­low irrigation, low tillage­high irrigation, high tillage­low irrigation, and high tillage­high irrigation (see Table 4.7). Low irrigation is defined as the absence of irrigation or manual irrigation, while high irrigation refers to use of mechanized (diesel or electric) pumps. Low tillage is defined as the use of bullocks in land preparation, while high tillage refers to the use of power tillers or tractors. Restoring Balance: Bangladesh's Rural Energy Realities 43 Table 4.7. Farm Household Distribution by Tillage/Irrigation Technologies Rice Cropping Season (%) Energy Technology Combinations Aus Aman Boro Overall Low tillage­Low irrigation 98.1 93.1 41.9 39.4 Low tillage­High irrigation 1.3 4.9 56.3 58.2 High tillage­Low irrigation 0.5 1.4 0.4 2.4* High tillage­High irrigation 0.1 0.5 1.4 * Source: BIDS Survey (2004). *Indicates merger of these two groups. Modern energy use is particularly relevant during the boro season and, to a considerable extent, in the aman season. Rural Bangladesh has only two major combinations of energy technology: low tillage­low irrigation and low tillage­high irrigation. Empirically, it is the difference between the first two categories that indicates the productivity-raising effect of modern energy, particularly during the boro season and in paddy cultivation, which accounts for most farmed land. The study found productivity differences by energy type for boro and aman paddy for owned land cultivated by the sample respondents (see Table 4.8). Farm productivity is sig- nificantly higher during the boro and aman seasons (22 percent and 17 percent, respectively). Thus, as is commonly known, irrigation provides a significant improvement in productivity. Distribution of Irrigation Gains by Land Ownership Although likely to have higher energy use, larger farms generally have lower land produc- tivity than smaller ones. To purge the confounding influence of such factors on land yield, this study conducted a regression analysis for the average gross yield of boro paddy land. Results of the analysis suggest that, on average, when the influence of such Table 4.8. Gross Productivity of Paddy factors as education of household head, by Tillage/Irrigation size of owned land, and regional varia- (Tk/decimal) tions are accounted for, up to 15 percent Rice Cropping improvement in productivity can be Season attributed to switching from low tillage­ Energy Technology Ladder Boro Aman low irrigation to the higher irrigation Low tillage­Low irrigation 118a,b 84a,b scenario. In absolute terms, the figure totals nearly Tk 1,800 per acre. For aman Low tillage­High irrigation 145a 96a paddy, by contrast, estimated improve- High tillage­Irrigation 140b 68b,a (low and high) ments are substantially smaller or may not be clearly discernible because irri- Source: BIDS Survey (2004). gation may be supplemental, depending Note: 1 decimal = 50 square yards; comparisons are for the same crop (within same column) on rainfall level in a particular location between technology levels. or farmer's land conditions. a Figures differ at 1 percent significance level. In reality, the gains may be even b Figures differ at 5 percent significance level. more substantial and inequitably dis- 44 World Bank Working Paper Table 4.9. Paddy Productivity in Boro Season by Land Ownership (gross value in Tk/decimal) Technology Land Ownership Observed Estimated (decimal)* Low-energy High-energy Low-energy High-energy 1­49 164 147 (­10) 140 149 (6) 50­249 141 145 (3) 123 144 (17) 250­500 96 146 (52) 99 138 (39) >500 80 141 (76) 89 130 (46) Average 118 145 (23) 107 141 (32) Source: BIDS Survey (2004). Note: Figures in parentheses represent percentage change over the low-energy situation. *1 decimal = 50 square yards. tributed. Table 4.9 illustrates the issue for boro paddy production. Both sets of observed and estimated yield figures for low-energy technology users exhibit the classic pattern of negative size­productivity relationship. But the observed yields for high-energy technology users are similar for every land-ownership class. The change in land productivity from energy use is likely masked by the influence of other factors. Although both observed and estimated yields fall with rising land ownership, the one high in energy use declines at a slower rate than the one low in use, indicating increasingly greater gains from higher farm technology (energy) use. The percentage differences in estimated yields for high energy use are somewhat less than those for the observed figures, but the patterns are similar. Furthermore, the average gain is 32 percent (Tk 3,400), nearly double that suggested initially by the regression equation. The conclusion is that farm households that move up the energy ladder enjoy a substan- tial gain in productivity and can provide irrigation or purchase irrigation water for their farms. About 50 percent of the average gain in transitioning from lower- to higher-energy technol- ogy is derived from switching to more intensive irrigation using mechanized (diesel or electric) irrigation pumps, with complementary factors accounting for the other 50 percent. Overall Income Gains: Moving Up the Energy Ladder Thus far, this chapter has examined some of the gains derived from using more modern energy for the specific areas of lighting, cooking, and irrigation. This section examines more general income gains that can be derived from moving up the energy ladder, many of which overlap with those already described. In short, the gains described in this section are a more general indicator of the value of the changes that result from improved cooking and living standards made possible by better ways of using energy. Cross-sectional Correlations of Electricity Use and Income This study survey measured household income separately for business and wage-related activities. More specifically, there are four measures of income, including agriculture Restoring Balance: Bangladesh's Rural Energy Realities 45 (farm-production activities), non-agriculture (entrepreneurial activities), agricultural wage labor, and non-agricultural wage labor. This section examines income levels for households with and without electricity, while the next one explores income gains from using more modern energy, controlling for many other factors. Overall, households with electricity have significantly higher incomes than those without electricity (see Table 4.10). This is generally true with regard to income from non-agricultural activities, except in divisions where irrigation is more prevalent, notably Rajshahi. In other divisions, the relationship may be reversed. Regarding agricultural wage labor, the relationship is counterintuitive, as it relates negatively to household electrification. Finally, non-agricultural wage labor is shown to have little relevance to electrification. Thus, it is mainly income from non-agricultural activities that has the most stable relationship with electrification. Given the low use of electricity for irrigation in Bangladesh, this is not an unexpected finding. Electricity is not the only modern energy source used in agriculture. The main fuel used for irrigation in Bangladesh is diesel. As discussed previously, rural households can gain substantially by moving up the energy ladder--whether switching from kerosene kupis to electric lamps, manual irrigation to electric pumpsets, or inefficient biomass stoves to improved models using cleaner-burning fuels. Similarly, farm productivity gains from switching to mechanized irrigation are estimated at 17 to 22 percent for any cropping season. Net Gains after Controlling for Other Variables It should be noted that correlation does not signify causality, and the tables presented thus far do not control for many factors. To quantify the effects of energy use on household income or consumption while controlling for other important factors, a regression technique is most appropriate. In this section, the results are presented only for the energy and outcome variables. The detailed models used for this analysis can be found in Annex 2. For this analysis, we divided household energy use into two relevant categories: (1) non-lighting and (2) lighting. Both types can be used for consumption and income generation. We further categorized energy use by major sources, namely, biomass, kerosene/ diesel, and electricity. Thus, the main outcome variables for the study are both farm and non-farm income (see Table 4.11). Perhaps surprisingly, in rural Bangladesh, non-farm income is higher than farm income. In addition to income, household consumption is considered an outcome.7 The study estimates the benefits of various forms of energy, especially modern energy, on household consumption (net of energy consumption) and income obtained from both farm and non-farm production. Table 4.12 shows the effects of energy use on household consumption and income by types of use (lighting and non-lighting) and source (biomass, kerosene/diesel, and electricity).8 The findings are interesting and in line with what one would expect from both farm and non-farm households moving up the energy ladder. To a certain extent, the findings reflect the efficiency with which a fuel is used. Biomass energy is generally burned in inefficient stoves, ovens, or boilers. Kerosene and diesel use generally has energy efficiency ratios of about one-third. At the end-use point, electricity is a very efficient form of energy. One would expect its effect to be higher than that of other fuels on both expenditure (consumption) and income. The results confirm that electricity 46 World Bank Working Paper Table 4.10. Household Income Differentials by Electrification Status (Tk per year) Electricity Division Income Source Connection Chittagong Dhaka Khulna Rajshahi All Divisions Agriculture No 31,776a 17,348 20,395b 19,751a 22,227 Yes 21,251a 20,169 28,898b 35,418a 23,980 Non-agriculture No 23,443a 24,729a 18,706a 13,538a 19,545a Yes 78,455a 58,236a 43,747a 29,504a 59,208a Agricultural wage labor No 4,821a 3,086a 4,606a 6,439a 4,868a Yes 2,240a 1,775a 1,168a 2,883a 2,085a Non-agricultural wage labor No 3,865 5,972 2,969 2,430 3,624 Yes 2,915 6,472 4,569 2,490 4,334 Total No 63,906a 51,135a 46,677a 42,159a 50,264a Yes 104,860a 86,652a 78,383a 70,296a 89,607a Source: BIDS Survey (2004). a Differences are statistically significant at 1 percent level or less. b Differences are statistically significant at 5 percent level. Restoring Balance: Bangladesh's Rural Energy Realities 47 has the highest level of impact on both Table 4.11. Summary Statistics of Outcomes and Energy Use Variables of energy expenditures and the equivalent Household Welfare Regressions measure of income. As might be expected, lighting Standard energy has a greater effect on non-farm Variable Mean Deviation income than farm income. Conversely, Outcome (Tk/month) non-lighting energy has a greater effect Consumption 3,581 2,449 on farm income than non-farm income. Farm income 2,229 2,702 Thus, lighting is more important for non- Non-farm income 2,889 4,997 farm income, while non-lighting energy, Total income 5,118 5,950 such as diesel, is more important for Monthly energy use farm income. After all, diesel can be Non-lighting (kgoe) 95.7 70.7 used for irrigation, which, as noted pre- Lighting (klm-hr) 115.6 267.7 viously, is strongly correlated with farm Biomass (kgoe) 90.1 65.7 productivity. Kerosene/diesel (kgoe) 3.5 9.5 The effects of different types of Electricity (kWh) 25.7 252.3 energy consumption (biomass, kerosene/ diesel, and electricity) are even more Source: BIDS Survey (2004). interesting. Biomass energy is the fuel most commonly used by households in Bangladesh. An increase of 1 kgoe of biomass energy increases monthly household income by about Tk 8, most of which can be attributed to farm income. Moving up the energy ladder yields greater income gains. Use of kerosene or diesel fuel significantly raises farm income, likely resulting from diesel-powered irrigation. Use of 1 kgoe (about 1 liter of fuel) increases farm income by approximately Tk 149, significantly higher than its cost. Of all fuels, electricity has the greatest effect; 1 kWh contributes to raising non-farm income by Tk 4.7 and total income by almost Tk 16, which is equivalent to more than Tk 185 for 1 kgoe, a gain of more than 20 times that achieved from biomass use. Much of the Table 4.12. Estimates of Household Welfare (Tk per month) Monthly Energy Use Consumption Farm Income Non-farm Income Total Income Model 1 Non-lighting (kgoe) 4.89** 9.61** 3.80 11.20** Lighting (klm-hr) 4.32** -0.19 9.68** 9.39** Model 2 Biomass (kgoe) 3.43** 13.32** -10.37 8.18** Kerosene/diesel (kgoe) -18.00 149.07* -151.61 -54.59 Electricity (kWh) 7.43** -2.17 4.72** 15.73** Source: BIDS Survey (2004). Note: Figures represent changes in outcome (consumption and income) due to one unit change in energy use. * = significance level of 10 percent. ** = significance level of 5 percent or stronger. 48 World Bank Working Paper non-farm income gain from electricity use can be attributed to lighting, the most common rural use of electricity. But one should not discount electricity's importance to farm income in rural Bangladesh, despite its limited use in irrigation. Income and other welfare gains are indicators of improved energy use by rural households.9 Given that biomass is used only for cooking, kerosene for both cooking and lighting, and electricity for lighting only, one can infer that household welfare would increase significantly by switching from biomass to kerosene for cooking and from kerosene to electricity for lighting. As previously discussed, real improvement in household income and consumption can be achieved by switching from biomass to kerosene or from kerosene to electricity. The evidence is fairly conclusive that switching to modern energy sources can increase a household's income and consumption both directly and indirectly. Increased income leads to greater consumption and more diversified investment, which, in turn, creates more income and household welfare. Summing Up This chapter has demonstrated that moving up the energy ladder can contribute significantly to rural income growth. Large welfare gains result when households switch from kerosene to electric lighting. Land productivity and agricultural income increase when farm households switch from manual to mechanized tillage and irrigation. The indirect link between electrification and education is also positive. Households with electricity have higher levels of education, although the causal direction is unclear. Children who live in households with electricity appear to study longer, thereby laying the foundation for longer-term, human-capital formation. Since mainly the poor lack electricity, a major concern is that children in poor households miss out on the educational benefits that electricity affords. Another concern is IAP, caused by the incomplete combustion of biomass for cooking and heating. Although the effects are not yet clearly understood, switching to cleaner-burning fuels, if affordable, and using improved stove designs, can mitigate the health risk to rural households. Modern energy's returns to household welfare and farm productivity are substantial, even after controlling for such factors as education, assets, and household location. Switching to modern energy and the modern use of biomass also contributes to the productivity of rural industries and microenterprises, the subject of the next chapter. Notes 1. The 2004 survey was conducted by the National Institute of Population Research and Training (NIPORT), Mitra and Associates and Measure/DSH+. 2. Colds, coughing, breathing problems, tuberculosis, headaches, eye problems, chest pains, and fever were compiled to give a broad idea of the respiratory problems associated with IAP. 3. In most localities of rural Bangladesh, the soil has low sand content, and mud walls and floors are frequently recoated to prevent cracking. The effective mud seal permits little ventilation, compared to other common construction materials (e.g., thatch or corrugated iron). 4. Health evidence is just reaching the point where people are realizing the long-term health consequences of IAP. Restoring Balance: Bangladesh's Rural Energy Realities 49 5. Farmers accounted for 61 percent of the survey sample; they were involved in some form of culti- vation during the 12 months preceding the survey. 6. For example, while data on crops grown and output, including by-products, have been requested from respondents on a disaggregated level, cost figures have been taken on an aggregated level. In such cases, total costs of fertilizer purchase may be accurate overall; but aggregate information on labor costs may have a significant margin of error. Unfortunately, collection of labor-related data is a detailed exercise; thus, gross, rather than net, returns have sometimes been used. Furthermore, labor use, which usually differs by tenure arrangement, adds to the complexity in determining net returns. 7. Household energy expenditure is excluded from this consumption figure. 8. Tables A2.2 and A2.3 (Annex 2) present complete regression results. These are marginal gains, which differ from the gross average gains calculated earlier. 9. The consumer or producer's surplus approach used to calculate consumption gain does not represent a real increase in household or farm income; rather, it estimates potential gain; that is, the savings a household could realize by switching from a lower-tier energy source to a higher one (Annex 3). CHAPTE R 5 Energy Consumption and Rural Production A round the globe, microenterprises have demonstrated their ability to facilitate economic growth and development. China's development of small- and medium- sized enterprises, for example, has contributed to the country's rapid economic advance. Such programs have been complemented by rural electrification and other rural infrastructure investments. Indeed, without the underpinnings of electricity and other development inputs, rural microenterprises would struggle. But simply introducing modern energy into the rural economy is unlikely to produce an economic miracle. Substantial evi- dence suggests that other supporting conditions are needed for businesses to grow and flourish. Under the right conditions, long-term growth can be sustained. This chapter charts the course of modern energy use in Bangladesh's rural micro- enterprises, exploring their energy-use characteristics and the effects of energy consumption on business growth and development. The rich data set collected by this study--including information on enterprise type, assets, and employment--reflects the need to conduct sep- arate surveys on rural household and growth-center microenterprises. A brief section on energy and farming highlights the importance of irrigation for productivity. Enterprise Types and Distribution Bangladesh's rural enterprises can be grouped into three location-based categories: growth center, village, and home. The survey findings reveal, perhaps surprisingly, that most home-based businesses center on manufacturing, while village and growth center In addition to the main authors, Rashid Faruqee and Hussain Samad were contributing coauthors of this chapter. 51 52 World Bank Working Paper Table 5.1. Microenterprise Distribution by Location and Activity Type Microenterprise Type Agriculture Services and Food Manufacturing Trading and Other All Location No. % No. % No. % No. % No. % Home 5 3.6 73 53.3 36 26.3 23 16.8 137 100 Village 63 18.4 12 3.5 208 60.8 59 17.3 342 100 Growth 197 10.9 191 10.6 1,116 62.0 297 16.5 1,801 100 center All 265 11.6 276 12.1 1,360 59.6 379 16.7 2,280 100 Source: BIDS Survey (2004). industries are involved more in trading. On closer examination, the reasons are clear. Many home businesses are run by women involved in such production activities as basket weaving and sewing. Outside the home, village enterprises, run mostly by men, are involved in store operations selling goods and services (see Table 5.1). Growth-Center Microenterprises Bangladesh has formally classified some 2,078 growth centers (large-sized markets).1 This study focused on rural enterprises randomly selected from 115 formal growth centers across 40 subdistricts (upazilas). Eighty percent of the selected growth centers were located 15 kilometers or more from the nearest district headquarters, while subdistrict centers were also situated at a considerable distance (about 15 percent of growth and upazila centers were indistinguishable). Most growth centers had access to roads in fair condition. Concentration of enterprises varied greatly by growth center. Some centers had no more than 15 enterprises, while others had as many as 808. All enterprises in the centers surveyed were self-standing (none included household-based enterprises). Seventy-five percent of the selected growth centers had been recently established, as evidenced by enterprise distribution by age of the growth centers. None of those established within one year of the survey involved manufacturing. Those established 5 to 10 years earlier had the highest concentration of such enterprises. The study found that, compared to other rural areas, growth centers enjoyed more extensive access to electricity. Nearly 94 percent of those studied had electricity; 76 percent obtained a connection from the rural cooperatives or PBSs (Palli Bidyut Samities), and 18 percent from the Bangladesh Power Development Board. Although only 38 percent had phone connections, nearly 67 percent had access to mobile phones. Types and Distribution Growth-center microenterprises typically focused on retail services. Trade and business (food and non-food) predominated, together comprising 61 percent of enterprises; these were followed by services (27 percent) and manufacturing (10 percent). About 38 percent of rural enterprises were linked to agriculture, mostly in the retail business. There was little regional variation (see Table 5.2). Restoring Balance: Bangladesh's Rural Energy Realities 53 Table 5.2. Distribution of Growth-Center Microenterprises by Region Establishments by Region Chittagong Dhaka Khulna Rajshahi Total Enterprise Type No. % No. % No. % No. % No. % Trade and business 153 33 160 33 142 34 145 31 600 33 (food) Trade and business 116 25 139 29 118 28 143 31 516 28 (non-food) Services 128 28 120 25 123 29 123 26 494 27 Manufacturing 54 11 58 12 34 8 45 9 191 10 Source: Data International Survey (2002). In the trade and business (food) category, groceries and general stores predominated, followed by tobacco (biri), betel quid (paan), and beetle nut shops. In the trade and business (non-food) category, the main establishments were traders of pharmaceutical and medical goods; cosmetics and toiletries; and textiles, clothing, hosiery, footwear, and leather goods. In the services category, tea stalls, restaurants, tailor shops, and beauty salons were the major businesses. The manufacturing category mainly included non-grain processing (e.g., jewelry, wood furniture, sawmills, and wood-processing units). Unlike other South Asian countries, where rice mills are common, Bangladesh mills rice either in households or using mobile threshing machines. The survey results compare favorably with those of other recent studies. For example, the National Private-sector Microenterprise Survey, conducted in 2003, revealed a similar pattern of micro-, small-, and medium-sized enterprises. It showed that 27 percent of enter- prises (excluding fisheries) were agriculture-based;2 while 14 percent were in manufacturing, compared to 11 percent in this survey. Since the 2003 survey included medium-sized enter- prises, trade and services sectors combined represented only about 50 percent of enterprises, while they predominated in this survey. This survey showed that 95 percent of rural enterprises have sole owner operators, while 5 percent are partner owned. Current owners initiated most of the businesses, signifying their recent establishment. Similarly, the 2003 survey found that 94 percent of enterprises had sole owners, of which 90 percent were owner operated. This survey's findings revealed that all establishments are owned and operated by men. The average operating day is 12 hours; unlike agricultural activities, there is no peak business season. Rural microenterprises are small, run primarily by owners (with an average of two employees). Even the size of manufacturing establishments is small (averaging only four employees). Similarly, the 2003 survey revealed an average of 3.3 workers for all types of rural enterprises. This survey shows that women are an extremely small percentage of employees (1.5 percent); by contrast, the 2003 survey reported that women represented about 9 percent of the work force. The value of enterprise assets is limited. Only one in six owns the land on which the operation was established. Nearly 67 percent have no assets; that is, the total value of assets, on average, is roughly equivalent to the annual revenue of the enterprise. Although the rural-enterprise composition of growth centers may not be representative of enterprises nationwide,3 it is nonetheless helpful to compare current findings with earlier 54 World Bank Working Paper data on Bangladesh and other countries. For example, Islam (2001) notes that eateries and tea stalls dominated earlier commercial activities in Bangladesh. This survey shows that such activities remain significant, but others have emerged, reflecting a diversification of business operations. Energy Use Analysis of energy-consumption patterns in Bangladesh's growth centers underscores the need to consider the context of energy-services delivery and the relative efficiency of the energy sources used. The survey confirms that firms use multiple energy sources, depending on their specific business needs (see Table 5.3). Nearly all firms use some form of electricity, mainly for lighting. If owners cannot access electricity from the cooperatives or national grid, they generate their own or purchase it from small firms. Because electricity is used more efficiently than other energy sources, its consumption appears lower (9 percent). Kerosene and diesel are major sources of backup lighting during power outages, while candles are also used. One surprising finding is that only 6 percent of firms--111 out of 1,801 enterprises-- use biomass energy, mainly for heat and steam. Charcoal is used by only 4 percent of firms to meet similar needs (see Table 5.4). Modern energy sources are used mainly for lighting and heating (including the prepa- ration of food for sale). Non-lighting uses include manufacturing and catering services and powering of small appliances. Use in drive-shaft power for productive purposes is not yet widespread. Table 5.3. Distribution of Energy Sources by Activity Type User Microenterprises (%) Users Trade and Trade and Energy (no. of Business Business Source firms) (food) (non-food) Services Manufacturing All* Biomass 111 3.8 0.2 15.6 5.2 6.2 Charcoal 73 0.0 0.4 6.3 20.9 4.0 Candle 1,034 58.2 64.1 52.6 49.2 57.4 Kerosene/diesel 986 59.8 39.7 59.9 66.0 54.7 Dry cell 196 11.2 13.6 8.9 7.8 10.9 Storage cell 57 2.5 4.3 4.0 0.0 3.2 Captive 343 19.5 21.7 18.0 13.1 19.0 electricity Grid electricity 1,472 79.0 86.6 79.6 82.7 81.7 Miscellaneous 56 3.2 4.1 2.0 1.6 3.1 None 26 2.7 0.8 1.0 0.5 1.4 Total 1,801 600.0 516.0 494.0 191.0 NA Source: Data International Survey (2002). * Sum of percentages is more than 100 because enterprises use multiple energy sources. Restoring Balance: Bangladesh's Rural Energy Realities 55 Table 5.4. Major Energy Sources by Use Type Microenterprises Use as Major Use as Major that Use Fuels Lighting Source Non-lighting Source No. of % of Total No. of % of Energy No. of % of Energy Energy Source Users Respondents Users Users Users Users Biomass 111 6 0 0 71 63 Charcoal 73 4 0 0 47 64 Candle 1,034 57 66 6 -- -- Kerosene/diesel 986 54 352 35 180 18 Generator electricity 343 19 122 35 0 0 Grid electricity 1,472 81 1,120 76 208 14 Source: Data International Survey (2002). Energy Costs For most growth-center microenterprises, energy is a significant but not major cost, averaging less than 5 percent of total operating expenses (from less than 2 percent in the food trade to about 10 percent for services) (see Table 5.5). As already illustrated, not all firms use all sources of energy for lighting and non-lighting purposes. If such a distinction is made, the proportion of energy costs rises appreciably, particularly for non-lighting uses. Although energy comprises a comparatively lesser share of total expenses, energy costs vary by energy intensity of the business. For example, retail shops incur small lighting costs, while diesel- or electricity-powered heavy machinery shops must budget more for energy. Table 5.5. Energy as a Percentage of Total Costs by Activity Type Microenterprise Type (%) No. of Trade Trade Energy Source Firms (food) (non-food) Services Manufacturing All Firms Biomass* 110 3.9 1.8 5.3 4.3 4.9 Charcoal 71 0 16.2 11.0 1.9 5.8 Candle 1,034 0.2 0.3 0.7 1.0 0.5 Kerosene/diesel 985 1.0 1.3 4.7 2.6 2.4 Dry cell 180 0.1 0.2 0.8 0.1 0.2 Storage cell 57 0.3 0.8 1.9 0 1.1 Generator 343 0.6 0.7 4.5 1.7 1.7 electricity Grid electricity 1,472 0.8 1.4 5.4 6.1 2.8 All sources 1,775 1.7 2.3 10.1 8.3 4.9 Source: Data International Survey (2002). * Biomass sources include fuelwood, sawdust, tree residue, straw, rice husk, jute sticks, bagasse, and other crop residue, animal dung, and briquette. 56 World Bank Working Paper Table 5.6. Price per Unit of Energy for Non-lighting Business Use Enterprise Type (Tk/unit) Energy Unit of No. of Trade Trade All Source Energy Users (food) (non-food) Services Manufacturing Enterprises Biomass* kgoe 71 3.1 1.3 3.3 3.9 3.3 klm-hr 0 -- -- -- -- -- Charcoal kgoe 47 -- 3.5 4.2 4.9 4.6 klm-hr 0 -- -- -- -- -- Kerosene/ kgoe 180 22.6 23.0 22.3 23.5 22.7 diesel klm-hr 301 1.3 1.0 1.4 1.5 1.4 Grid kgoe 208 55.9 86.4 66.1 53.4 66.3 electricity klm-hr 1,173 0.04 0.02 0.03 0.03 .03 Candles klm-hr 72 48.2 41.8 39.2 37.9 41.3 Source: Data International Survey (2002). * Biomass sources are fuelwood, sawdust, tree residue, straw, rice husk, jute sticks, bagasse, and other crop residue, animal dung, and briquette. Explaining such cost variations requires examining the level and effects of relative energy pricing for various fuels used by rural enterprises. Energy cost is relative to what businesses pay for it. Predictably, based on standard energy units, the price of electricity is higher than that of all other energy types, mainly because of the higher value of the energy services provided. Measured by energy content, biomass is the cheapest source, while electricity is the most expensive. Although the price of charcoal is low, its use is specialized (e.g., making jewelry and tobacco [biri]). Con- versely, for lighting services (measured in kilolumen-hours), electricity is the cheapest source (Tk 0.003 per klm-hr versus Tk 1.4 for kerosene and Tk 41.3 for candles). Indeed, for the same lighting service, kerosene is about 45 times costlier than grid-based electricity (see Table 5.6). To summarize, patterns of energy choice are, in large part, determined by the relative fuel price for the desired end use. For heating, biomass is the least expensive fuel. For lighting, grid electricity is by far the cheapest source. Thus, it is not surprising that most rural microenterprises switch to grid electricity when it becomes available in their respective communities and otherwise generate their own electricity or purchase it locally from small generator firms. Entrepreneurial Attitudes The study survey asked business owners and operators a variety of energy-related atti- tudinal questions. Findings show that business entrepreneurs consider electricity the best source of lighting, but also perceive that reliability of supply is a significant prob- lem. Few respondents consider kerosene superior to electricity for lighting. Most agree that electricity provides better illumination than kerosene and makes reading easier. They also perceive that grid electricity is superior to batteries for powering television (see Table 5.7). Restoring Balance: Bangladesh's Rural Energy Realities 57 Table 5.7. Perceived Advantages and Disadvantages of Electricity Use Response Choices (%) Strongly Indifferent/ Strongly Survey Statement Agree Agree Neutral Disagree Disagree Electricity provides better 70.7 28.9 0.1 0.2 0.0 illumination than kerosene oil (n = 1,801) Reading is easier with 64.4 34.8 0.8 0.1 0.0 electric lamp than kerosene lamp (n = 1,801) It is [more] advantageous to 60.9 35.1 3.5 0.2 0.3 run TV by electricity than battery (n = 1,766) Source: Data International Survey (2002). Entrepreneurs perceive irregular electricity supply and voltage fluctuations as a negative use factor. Some 67 percent of respondents are convinced that PBS power distribution and management quality have declined in recent years. Indeed, nearly 75 percent consider fatal accidents as a disadvantage of use. About 67 percent view electricity as expensive, and most consider it an unattractive option for cooking, probably because of its high cost compared to biomass fuel. However, nearly all respondents agree that the advantages of using electricity far outweigh the disadvantages. Enterprises without electricity were asked several questions related to the costs and benefits of electricity use. There was no consensus on the cost of obtaining a business connection: About 50 percent did not consider prevailing rates exorbitant, 30 percent thought they were too high, and about 19 percent were indifferent. This finding is consistent with respondents' reactions to a similar statement on affordability of connection. About 59 percent of non-users said that, if they were to obtain access, the monthly payments would not be too high; about 44 percent perceived it as unaffordable. In short, despite the overwhelming perception regarding net gains, most respondents believed the advantages of electricity outweighed the problems. With regard to fuelwood, the opinions expressed varied markedly, reflecting the widespread problem of deforestation and thus uneven resource availability. Because of scarcity and moderate shortages, about 50 percent considered fuelwood expensive. At the same time, many rural entrepreneurs perceived fuelwood as a readily available resource. More than 90 percent expressed concern with regard to deforestation. Rural entrepreneurs preferred fuelwood to straw, dung, and other biofuels, which were used sparingly. But about 85 percent of respondents were aware that the smoke emitted from fuelwood could cause respiratory problems, and some 50 percent agreed that it may cause other health problems (see Table 5.8). Respondents' opinions differed on the use of kerosene and liquefied petroleum gas (LPG). Kerosene, although readily available, was viewed as a health hazard. Rural entrepreneurs perceived its price as relatively high, but were unclear why. More than 55 percent considered LPG a good source of cooking energy. About 38 percent considered LPG as very expensive, but few respondents were current users, reflecting the perception of high price versus 58 World Bank Working Paper Table 5.8. Perceived Advantages and Disadvantages of Fuelwood Use Response Choices (%) Strongly Indifferent/ Strongly Survey Statement Agree Agree Neutral Disagree Disagree Fuelwood is readily 31.2 33.3 0.3 17.9 17.3 available to collect (n = 1,801) Fuelwood can be obtained 30.3 32.6 2.3 15.1 19.8 free of cost from own trees (n = 1,801) Fuelwood creates smoke 40.1 45.1 5.1 4.3 5.3 that creates breathing problems (n = 1,801) Cooking with fuelwood 25.0 33.4 10.5 16.6 14.5 creates health hazards (n = 1,750) Fuelwood is very expensive 28.3 32.5 1.6 15.3 22.3 for cooking (n = 1,798) Source: Data International Survey (2002). actual cost. Many answers to LPG-related questions were indifferent, suggesting that most respondents were unaware of the fuel (see Table 5.9). The picture that emerges from this survey is that most growth-center microenterprises in rural Bangladesh are not energy intensive. They consist mainly of small retail stores and shops, with some manufacturing. The main energy use is lighting, for which virtually all businesses use electricity. If a particular growth center does not yet have electricity, business owners organize their supplies, either by buying their own generator or purchasing small amounts from a local firm. Small agricultural industries and restaurants that require energy for heat rely mainly on biomass or kerosene. Table 5.9. Perceived Advantages and Disadvantages of Kerosene Use Response Choices (%) Strongly Indifferent/ Strongly Survey Statement Agree Agree Neutral Disagree Disagree Cooking with kerosene 17.2 24.5 4.1 30.6 23.6 is easy (n = 1,801) Kerosene is not expensive 14.1 21.6 2.2 38.6 23.4 for lighting (n = 1,801) Kerosene is very expensive 59.6 30.5 3.8 3.4 2.6 for cooking (n = 1,801) Cooking with kerosene 46.8 35.7 11.2 3.1 3.2 creates health hazards (n = 1,801) Source: Data International Survey (2002). Restoring Balance: Bangladesh's Rural Energy Realities 59 Home and Village Enterprises The survey showed that many rural microenterprises are located outside the growth centers. Out of 2,400 households, 137 had home-based enterprises; in addition, 342 village-based enterprises were located outside rural households. Types and Distribution As already mentioned, most home-based businesses are women run and involve manufac- turing (basket weaving and sewing), while village industries, run mainly by men, focus more on trading (selling of goods and services). The age of home and village microenterprises is spread over a number of years. More than 25 percent have been in operation for a decade or longer, although interesting differences are evident between categories. For example, about 40 percent of home-based enterprises are more than 10 years old, twice the proportion of village-based ones. Few home-based enterprises have been in operation for one year or less, while more than 17 percent of village enterprises have been recently established (see Table 5.10). Home-based manufacturing enterprises likely involve handicraft activities carried down by families over generations. Some 50 percent of home- and village-based microenterprises are run by only one person, although differences between categories are apparent. During peak periods of operation, about 31 percent of home enterprises employ only one person, compared to some 58 percent of village enterprises. In fact, home enterprises are more evenly distributed in terms of size of operation (proxied by number of workers during peak periods). This finding reinforces the distinction between manufacturing centered mainly in home enterprises and trading in village-based ones (see Table 5.11). Energy Use and Costs The survey findings show that more than 70 percent of home enterprises and virtually all village enterprises--most of which operate during extended evening hours--use energy for lighting. Beyond lighting, however, these small handicraft and retail firms use little energy. About 67 percent of home enterprises and 64 percent of village ones use no energy Table 5.10. Microenterprise Distribution by Location and Length of Operation Years in Home-based Village-based Total Operation No. % No. % No. % <1 11 8.1 58 17.2 69 14.2 1­3 26 19.3 91 27.0 117 24.8 3­5 14 10.4 48 14.2 62 13.1 5­10 29 21.5 75 22.3 104 22.0 >10 55 40.7 65 19.3 120 25.4 Total 135 100.0 337 100.0 472 100.0 Source: BIDS Survey (2004). 60 World Bank Working Paper Table 5.11. Microenterprise Distribution by Number of Employees and Location Workers in Peak Home Village All Firms Periods (no.) No. % No. % No. % 1 40 30.7 182 58.3 222 50.2 2 41 31.5 93 29.6 134 30.3 3­5 43 33.0 34 10.9 77 17.4 >5 6 4.0 3 0.9 9 2.0 Total 130 100 312 100 442 100 Source: BIDS Survey (2004). for non-lighting purposes. Some report using no energy types for either lighting or non- lighting purposes, implying that their businesses involve manual labor and operate mainly during the daytime. Compared to village enterprises, home-based businesses use significantly less energy; on average, home enterprises consume 332 kgoe per year, compared to 608 kgoe for village- based ones (including non-users). Predictably, home businesses depend more on biomass than do village enterprises. But surprisingly, within the biomass portion of the energy basket, home enterprises depend more on fuelwood, while village enterprises use non-fuelwood biomass; cow dung accounts for 8 percent of village enterprises' energy basket, compared to only 1 percent for home enterprises. Similarly, village enterprises derive 49 percent of their energy from crop residue, compared to only 26 percent for home enterprises (see Figure 5.1). Figure 5.1. Comparison of Energy Use Structure by Enterprise Location Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 61 A major reason is that manufacturing businesses, mostly home-based, require better heat control and are thus more likely to use superior forms of biomass. A portion of the needed resources may be collected from the owner's homestead, implying an easier supply for home-based, versus village, enterprises. Findings also reveal differences in other portions of the energy basket. With regard to liquid fuels, home enterprises use more diesel than kerosene; conversely, village enterprises use more kerosene than diesel. Home enterprises, on average, consume only about 6 kgoe of kerosene, compared to nearly 68 kgoe for village enterprises. With regard to electricity, village enterprises use more than 30 kgoe, compared to only 3.9 kgoe for home enterprises (see Figure 5.1). Clearly, trade-oriented village enterprises require more lighting services than do home-based businesses. Given that non-lighting energy use predominates in home and village-based enterprises-- as evidenced by the critical role biomass plays in the energy balance of both--it is of inter- est to compare the output and costs of energy sources for non-lighting and lighting uses (see Table 5.12). On average, village enterprises consume more energy per enterprise for non-lighting purposes than do home enterprises (657 versus 372 kgoe per year). Indeed, for each enterprise type, village enterprises consume higher levels of energy. Furthermore, the lighting services consumed by village enterprises are more than six times that of home- based enterprises, suggesting the importance of lighting for village businesses and the small size of home enterprises, some of which lack access to electricity. The price of energy per kilograms of oil equivalent varies markedly by fuel source. For non-lighting uses, biomass is by far the least expensive source, which accounts for its dominance as a heating fuel. Because village enterprises are larger than home-based ones, they use significantly more biomass energy for non-lighting uses. Table 5.12. Comparison of Energy Output and Cost by Use and Location Lighting Non-lighting Amount Price Amount Price Energy Source (klm-hr/year) (Tk/klm-hr) (kgoe/year) (Tk/kgoe) Home Biomass 0.00 -- 317.5 4.34 Candle 0.32 43.33 0 -- Kerosene or diesel 23.40 79.15 40.7 29.11 Grid electricity 461.50 0.77 14.2 185.62 All sources 474.20 17.92 372.4 27.21 Village Biomass 0 -- 546.2 3.68 Candle 1.1 54.68 0 -- Kerosene or diesel 37.7 17.11 95.1 32.08 Grid electricity 2,877.0 0.27 16.0 55.52 All sources 2,915.8 10.50 657.2 29.20 Source: BIDS Survey (2004). 62 World Bank Working Paper The price for lighting is similar to the results presented for rural households in Chapter 3. For home enterprises, the average price of electric lighting is Tk 0.27 per klm-hr and more than Tk 40 per klm-hr for candles and kerosene; findings show similar prices for village-based enterprises. Given the wide disparity in the price of lighting, it is not surprising that businesses use more electricity when they need lighting; electric lights are used primarily, with kerosene lamps and candles as backup sources. The benefits of switching to a higher-tier source along the energy ladder extend beyond profitability, particularly for lighting services. The cost of one unit of lighting service from electricity is far cheaper than that from kerosene. For home and village-based enterprises, the respective costs of electric lighting are Tk 0.77 and 0.27 per klm-hr. The corresponding costs for kerosene lighting are Tk 30.40 and 17.11 (nearly 40 and 63 times greater). Clearly, a major reason for the higher profitability of electricity-based enterprises involves such cost differences for lighting. Enterprise Energy Demand and Profitability: A Quantitative Analysis Rural microenterprises, like rural households, are influenced by a range of enterprise- and community-level factors when selecting from among alternate energy sources. Their choices, in turn, eventually affect the enterprise outcome; that is, revenue and profit. Enterprise fac- tors may include years that a business has been in operation, assets, and types of activi- ties; community factors may include prices of alternate energy sources and infrastructure variables. Table 5.13 provides sum- mary statistics of energy variables and selected influencing variables, Table 5.13. Summary Statistics of Outcomes and while Table 5.14 presents regression Explanatory Variables of Enterprise Energy Demand Regressions results of energy demand.4 An enterprise's non-land assets Variable Mean have a significant positive effect Enterprise energy demand on consumption of biomass (fuel- Fuelwood (kg/month) 73.25 wood), diesel fuel, and electricity. Kerosene (liter/month) 164.72 Although the price of fuelwood has Diesel (liter/month) 60.39 no effect on any type of energy Electricity (kWh/month) 218.58 source, the prices of diesel and elec- Explanatory tricity significantly affect energy Enterprise assets demand. The price of diesel posi- Land (acres) 26.34 tively affects electricity consump- Non-land (Tk 10,000) 1.25 tion, thereby showing that diesel and Village/growth center price electricity are substitutes in enter- Fuelwood (Tk/kg) 1.36 prise consumption. If this statement is true, however, a higher electricity Kerosene (Tk/liter) 19.71 price should lead to greater diesel Diesel (Tk/liter) 21.39 consumption. But results show a Electricity (Tk/kWh) 4.50 negative effect, which is possible if If the village/growth center has electricity 0.89 the negative income effect caused Source: Data International Survey (2002) and by the higher price of electricity BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 63 Table 5.14. Estimates of Enterprise Energy Demand (N = 2,290) Energy Demand (monthly) Explanatory Variable Fuelwood Kerosene Diesel Electricity Enterprise assets Land (acres) -0.15** 0.04 -0.04* 0.11 Non-land (Tk 10,000) 3.99** 1.14 6.12** 25.51** Village/growth center price Fuelwood (Tk/kg) 5.25 1.17 0.93 -10.46 Kerosene (Tk/liter) 11.43* -1.03 -3.76 -14.56 Diesel (Tk/liter) -4.79 -11.17 3.29 23.38* Electricity (Tk/kWh) -6.07 -15.87 -13.26* 25.68 If village/growth center has electricity 10.47 -26.91 -18.19** 418.72** Source: Data International Survey (2002) and BIDS Survey (2004). Note: * = significance level of 10 percent; ** = significance level of 5 percent or stronger. Figures represent changes in energy demand caused by unit changes in explanatory variables. outweighs the positive relative price effect of electricity on demand for diesel. Thus, diesel and electricity remain substitutes in energy use by rural enterprises. The demand for kerosene, which appears unaffected by these factors, is included here for completeness (see Table 5.14). Given the demand for alternate types of energy in non-farm production, one can estimate the effects of energy use on enterprise revenue and profit. Energy use is expressed in taka value adjusted by village and growth-center price indices. Table 5.15 provides summary statistics and Table 5.16 regression results.5 As Table 5.16 shows, energy con- sumption positively affects enter- prise profit. One taka worth of energy Table 5.15. Summary Statistics of Outcomes and consumption increases enterprise Energy Use Variables of Enterprise profit by more than half a taka. Profitability Regressions (N = 2,290) When the effect is differentiated by energy source, one discovers that Variable (Tk/year) Mean kerosene/diesel consumption has no Outcome effect on either profit or revenue, Revenue 555,817.6 but biomass and electricity have a Profit 72,590.8 significant positive effect, especially Energy on profit. One taka spent on bio- Total use* 6,674.82 mass consumption increases profit Biomass 1,129.62 by Tk 1.22; a similar amount spent Kerosene 2,701.26 on electricity increases it by Tk 1.35. Electricity 2,622.92 Thus, the effect is slightly higher for Source: Data International Survey (2002). electricity than for biomass as an * Includes the cost of miscellaneous sources, in addition energy source in enterprise income to the three major ones. generation; that is, enterprises would 64 World Bank Working Paper gain by switching from biomass and Table 5.16. Effects of 1 Tk Increased Energy diesel to electricity as an energy source Use on Enterprise Outcome in production. (annual Tk per enterprise) The conclusion is that use of (N = 2,290) either high-quality biomass or pur- Energy Source (Tk/yr)a Revenue Profit chased electricity dramatically affects Model 1 the profitability of rural Bangladeshi enterprises. To reiterate, biomass is Total energy use 0.86 0.55** usually the least expensive fuel and Model 2 is used primarily for heating, while Energy use electricity is used for lighting and Biomass 5.72* 1.22** other purposes. Thus, energy source Kerosene -3.82 -0.27 matters in the profitability of rural Electricity 4.11 1.35** businesses. Source: Data International Survey (2002). Note: * = significance level of 10 percent; ** = signifi- cance level of 5 percent or stronger. Figures represent Energy for Farming changes in outcome (consumption and income) caused by unit changes in explanatory variables. Farming is a critical energy use for a Total energy cost used in Model 1 is slightly higher rural households. The country's than the sum of individual energy costs used in Model 2 potential for increased irrigation, because it includes other miscellaneous sources. though not as developed as that of other South Asian countries, is significant. The most obvious energy inputs are diesel engines and electric motors for pumping irrigation water. In addition, a variety of small machines, including power tillers, are used increasingly in land preparation, alleviating much of the drudgery associated with such work. Farming Stages and Energy Inputs As discussed in Chapter 4, Bangladesh is characterized by three rice-cropping seasons: (1) aus (rainfed summer rice planted in March­April), (2) aman (rice planted in July­ September and harvested in November­December), and (3) boro (irrigated winter rice planted in December­February and harvested in April­June).6 Aus requires little energy for tillage (except bullock power) and irrigation; boro necessitates energy for both tillage and irrigation, and aman falls in between. The major rice-farming stages involve land preparation, irrigation, harvesting, and transport to market.7 During land preparation, bullock-driven ploughs or power tillers/ tractors are used. While bullock power may be quantified, Bangladesh has little relevant data available. Power tillers/tractors require petroleum fuel (mainly diesel). Irrigation, the next major stage, overlaps with land preparation; that is, when rice is transplanted, the land must be thoroughly flooded and puddled. Irrigation is also required at later stages of rice-growing. The capacity to irrigate land is limited to manual implements (e.g., swing baskets and hand tube wells). Mechanized methods that depend on modern energy use far costlier, but more efficient, equipment for large tracts. The three basic equipment pieces are low-lift pumps for surface water irrigation, surface tube wells (including deep-set shallow tube wells), and deep tube wells. These may be run with either a diesel- or electricity-operated Restoring Balance: Bangladesh's Rural Energy Realities 65 motor. Low-lift pumps are owned either privately or collectively. Surface tube wells are privately owned; while deep tube wells are usually owned collectively (for reasons of higher cost, as well as custom). After irrigation are harvesting and transport. Most harvesting is done manually. Threshing, traditionally done manually, depends more on diesel- or electricity-powered machines. Husking increasingly uses machines that run on diesel and occasionally kerosene. Crops may be transported to market by bullock or by using mechanized methods that depend on petroleum fuels. Land Tillage Methods Over the past decade, acreage under mechanized tillage has risen significantly. In 1996, 4.2 million acres--20 to 25 percent of the country's cultivated land--was tilled using diesel- or petroleum-powered tillers or tractors. When the 2004 household survey was conducted, that proportion had risen to more than 66 percent, although regional variations were substantial (see Table 5.17).8 Although use of power tillers and tractors is common, ownership is not. In the study sample, only 35 farmers owned mechanized tillage equipment, while 1,068 rented it; another 186 households rented draft bullocks, while the remainder used their own bullocks. Irrigation Methods Mechanized irrigation using modern energy is of growing importance in rural Bangladesh. In 2002­2003, 73 percent of the 11.68 million acres irrigated used mechanized tube wells; another 17 percent used power pumps, while 3 percent used canals (or in combination with power pumps) and the remainder traditional methods (BBS 2005). Although most mechanized irrigation equipment is diesel-driven, the percentage division between diesel and electricity is unclear. In 2005, the country had 7.54 million mechanized irrigation sets (including low-lift pumps, shallow tube wells, deep-set shallow tube wells, and deep tube wells) (BBS 2005). According to the Rural Electrification Board, in February 2006, irrigation connections totaled slightly more than 195,000.9 Even assuming that one connection could serve more than one set (which is seldom the case), electrically driven irrigation sets are uncommon. The village survey found that 68 percent of villages used mechanized irrigation methods. Of these, 70 percent used diesel power, 10 percent used electricity, and 3 percent used both (see Table 5.18). Table 5.17. Comparison of Tillage Methods by Division Division (%) Tillage Method Chittagong Dhaka Khulna Rajshahi Total Power tiller/tractor 45.9 75.3 61.5 86.6 66.1 Draft bullock 54.1 24.7 38.5 13.4 33.9 Source: Village censuses (2004). 66 World Bank Working Paper Table 5.18. Irrigation Method and Energy Type by Division Division (%) Method or Type Chittagong Dhaka Khulna Rajshahi All Divisions Irrigation Mechanized 52.0 78.5 53.4 91.5 68.1 Hand tube well 1.1 1.2 0.9 0.8 1.0 Indigenous method 10.1 15.0 19.8 1.1 11.2 None 36.8 5.3 25.9 6.6 19.7 Energy Diesel 50.0 93.3 66.7 76.7 70.4 Electricity 19.2 6.7 -- 20.0 10.4 Diesel and electricity 2.6 -- 7.4 3.3 3.2 Not applicable 34.2 -- 25.9 -- 16.0 Source: BIDS Survey (2004). Of the 1,459 farm households surveyed, only 62 percent irrigated their land (see Table 5.19). Of these, 97 percent used mechanized means. Only 130, or 9 percent of all respondents, owned mechanized irrigation sets.10 Ownership of mechanized pumps was concentrated in Rajshahi--77 out of 130 farming households--where 18 percent of such households owned modern irrigation equipment. But in Dhaka, half of non-farmers owned irrigation sets. Fuel Choice for Irrigation As noted previously, mechanized irrigation equipment may be powered by electricity or diesel. According to available information, only 112 of current farm households use their machines for land under their control (i.e., self-cultivated owned land and other land under lease agreement or other form of user rights). Of these, 103 are operated by diesel and 9 by electricity (8 of which are in Rajshahi). It is also noteworthy that about 70 percent of diesel-pump users are in villages without electricity. Average electricity consumption per farm and the costs borne by farmers indicate an expense of nearly Tk 2.8 per kWh, far lower than for domestic purposes (Tk 7.7 per kWh in Rajshahi and Tk 5.5 for all divisions). Average Table 5.19. Farmer Irrigation Method by Division Division Chittagong Dhaka Khulna Rajshahi Total Method No. % No. % No. % No. % No. % Manual 1 0.3 12 3.7 9 2.8 1 0.2 23 1.6 Mechanized 159 40.3 203 63.2 158 48.5 356 85.4 876 60.0 All 160 40.5 215 67.0 167 51.2 357 85.6 899 61.6 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 67 Table 5.20. Energy Consumption in Farming (average owner-user households per year) Electricity Diesel Consumption Cost (Tk) Consumption Cost (Tk) Yearly Per Yearly Per Equipment Type kWh Households Yearly kWh liters Households Yearly liter Irrigation pump 3,643.47 9 9,762 2.8 347.95 103 7,056 20.5 Power tiller/ -- -- -- -- 238.24 35 4,799 20.2 tractor Thresher -- -- -- -- 78.13 17 1,698 20.8 Source: BIDS Survey (2004). diesel consumption per farm and farmer costs are slightly more than Tk 20 per liter. The cost per liter is similar across types of use, but varies slightly by division (see Table 5.20). The study found that many farmers used water from mechanized irrigation equipment, whether owned or leased. The percentages were higher for farmers owning pumps, reflecting the active market for purchased irrigation water in rural areas. As expected, Rajshahi had the highest proportion of farmer ownership of irrigation equipment. On average, a farm household required about 348 liters of diesel per year for irrigation, reflecting an annual expense of nearly Tk 7,056. Thus, irrigation is extremely important for farm productivity, as well as profitability from the sale of irrigation water to neighbors. Clearly, farm households that could benefit from electricity-powered irrigation have not used electricity as extensively as they might, perhaps, in part, because of their remote location. Further research is needed to identify the factors that have prevented farmers from adopting electricity-driven pumps, particularly given the difficulty of moving large diesel pumps between fields and deciding whether government-subsidized electricity for irrigation is a sound policy. Summing Up In rural Bangladesh, microenterprise entrepreneurs own and operate a range of businesses, from trade and commercial services to small-scale manufacturing. Although small, these enterprises represent a substantial increase in rural people's incomes. The three types of microenterprises examined in this chapter--growth-center, village, and home-based--use a wide array of energy sources. All rural businesses in the growth centers use electricity, either from the grid or local generators; and most have electric lighting. By contrast, not all free-standing village and home-based enterprises have access to electricity; those that do may choose not to use electricity, even for lighting. Such enterprises depend heavily on biomass energy, mainly for heating and manufacturing needs. Though inconvenient, biomass is the least expensive heating fuel, a major factor in explaining rural businesses' continued reliance on it. Another factor is that most rural areas have little or no access to LPG.11 Bangladeshi entrepreneurs perceive that biomass is inexpensive compared to other fuels and that it can be collected from the environment. At the same time, they are aware 68 World Bank Working Paper of the health problems associated with traditional fuelwood use. They also agree that local biomass collection is becoming more difficult. Increasingly, they view traditional lighting sources as inefficient and of poor quality. The lighting analysis confirms that candles and kerosene lamps are many times more expensive than electric lighting. Microenterprises with electricity enjoy benefits and cost advantages related to lighting and profitability. Indeed, the average cost of electric lighting can be 60 times less expensive than kerosene, the next cheapest source. Unreliability of electricity supply, however, forces businesses to depend on kerosene and diesel as backup lighting sources. The survey findings indicate scope for improvement in the energy-use patterns of rural Bangladeshi microenterprises.12 Analysis of the relationship between modern energy sources, as defined by the energy ladder concept, indicates that rural businesses that use modern energy generally have more revenue and are more profitable than those that rely on traditional energy sources. Thus, a viable rural energy strategy for Bangladesh must consider the best ways to promote modern energy, as well as the modern use of biomass. To this end, it is critical to examine the institutional underpinnings of energy distribution, which are the subject of the next chapter. Notes 1. The Government of Bangladesh, through its Engineering Department and local government, maintains a regularly updated list of rural growth centers. 2. If fisheries had been included, agriculture would have comprised 32 percent, compared to 38 percent in the current survey. 3. Limiting business coverage to the growth centers means that many common, larger-sized enterprises (e.g., brick kilns, cold storage, fisheries, poultry, and hand looms), some of which are energy intensive, are not captured. 4. Table A2.4 (see Annex 2) presents the complete regression for energy demand. 5. Tables A2.5 and A2.6 (see Annex 2) provide complete regressions. 6. Details are available from the International Rice Research Institute (www.irri.org/irrc/weeds/ closing.asp). 7. To simplify data collection and analysis, this study emphasized farmers who directly manage their own land. Thus, analysis and discussion on energy use and its characteristics are tentative, as key data may be lacking. More definitive answers call for more in-depth studies. 8. In 2005, a large-scale survey found that, since 1996, the number of power tillers and tractors had risen from slightly more than 150,000 to 711,000 (BBS 2005). 9. Details are available at www.reb.gov.bd. 10. Several households owned more than one piece of equipment. 11. Generally unavailable in rural areas, LPG uses 60 percent of cooking energy, compared to only 15 percent for traditional biomass stoves. 12. Although energy for non-lighting use is limited, improvements are possible through the use of charcoal and more efficient kerosene stoves. CHAPTE R 6 Institutional Framework for Rural Energy Supply P romoting better rural energy policies in Bangladesh is inextricably linked to national poverty reduction and rural development. Extending both grid and off-grid elec- tricity systems improves rural livelihoods and enables delivery of critical human services, from clean water and sanitation to health and education. The results of this study demonstrate that access to improved biomass stoves, electric lighting, and mechanized tillage and irrigation creates the enabling conditions for informal and small-enterprise sectors to grow, thereby empowering the poor.1 Rural energy must be viewed as a basic input to the rural economy, in line with its role in rural productivity and income generation. Achieving a supply-and-demand balance is critical at all levels.2 Given rural residents' heavy reliance on traditional biomass energy, an urgent priority must be to increase biomass supply. To achieve a sustainable supply, biomass must be used in more efficient ways that mitigate damage to human and global environ- mental health. Complementing these efforts, more diversified modern energy sources must be made available to fulfill unmet rural household and business demand. To this end, appropriate pricing policies are vital: They must be market based to ensure that suppliers can sustain the higher cost of rural operations,3 yet remain affordable to the poor.4 Institutional Challenges and Potential To advance the development goals of poverty reduction and human development, the institutions responsible for delivery of modern energy services must be responsive to the unique needs of rural populations. The inefficiencies associated with Bangladesh's current In addition to the main authors, M. Iqbal, M. Khaliquzzaman and Grayson Heffner were contribut- ing coauthors of this chapter. 69 70 World Bank Working Paper energy-sector providers, which have shifted financial burden to the government and rural consumers, call for more effective institutional arrangements and coordination. This chapter reviews the current structure of Bangladesh's energy sectors, highlighting past institutional and policy deficiencies and opportunities for building a more integrated approach to meeting current and future rural energy demand. Biomass Sector This study suggests that Bangladesh's heavy reliance on the unsustainable use of traditional biomass, especially fuelwood, represents a significant opportunity cost and serious health risk for millions of the rural poor. Compounding their diminished quality of life is the accelerated commercialization of all forms of biomass--including low-quality tree leaves and grass.5 This section highlights lessons from past and ongoing efforts to improve biomass energy efficiency to increase rural productivity and quality of life. Improved Biomass Stoves. From the late 1980s to 2001, the Institute of Fuel Research and Development promoted improved cookstoves through several donor-supported projects. Although an estimated 400,000 improved stoves were distributed, commercialization failed, in large part, because projects focused on targets and subsidies rather than monitoring. Today the potential market for improved stoves is enormous. Better-designed stoves could save millions of days in lost productivity, decrease early mortality, and save energy. But to succeed, programs must promote stoves that people want. Thus, future institutional support should focus on stove design, manufacture, and testing, as well as training and quality monitoring. Biogas Production. Basic biogas technology holds special promise for rural Bangladesh. Biogas can be harnessed for cooking, lighting, and mechanized irrigation through village- based production and distribution networks. Such arrangements can enhance rural residents' self-sufficiency during periods of load shedding and diesel shortages. Currently, tens of thousands of household- and village-level biogas plants are in place throughout the country. Over the past 30 years, an array of government agencies, university research institutions, and non-governmental organizations (NGOs) have participated in the development, piloting, and implementation of household- and enterprise-scale biogas programs. Key entities are: Rural Electrification Board Local Government Engineering Department Bangladesh Council for Scientific and Industrial Research and Institute of Fuel Research and Development (a Council agency) Renewable Energy Research Centre (Dhaka University) Bangladesh University of Engineering and Technology Grameen Shakti Bangladesh Rural Development Board Bangladesh Rural Advancement Committee Proshika Although large-scale plant production appears economically and financially feasible, several key issues must be addressed. First, it is unclear whether the subsidy culture that Restoring Balance: Bangladesh's Rural Energy Realities 71 characterized past programs will create future difficulties. Second, it is likely that biogas plants will benefit better-off households more than the poor (Ghimire 2005). Third, the number of potential plants is unknown.6 Power Sector Bangladesh's power sector has lagged overall economic growth, resulting in chronic unmet energy demand. Only 40 percent of the total population and less than 30 percent of rural residents have access to electricity. Low connection rates and unreliable power supply result in fewer opportunities for economic development or improved rural livelihoods. Though rural households grasp the benefits of electricity, they perceive it as expensive, in part, because of frequent power outages. As Figure 6.1 illustrates, total effective power-generation capacity has fallen below electricity demand, resulting in extensive daily load shedding. In turn, load shedding has obscured the reality of unmet demand. Over the next nine years, electricity demand is expected to double (at 8 percent annual growth). The investment required to generate sufficient capacity to accommodate future demand growth is immense. In response to this challenge, the state-owned Bangladesh Power Development Board has put forward an ambitious expansion plan calling for installation of an additional 4,210 MW of generation capacity by 2010. The plan, which relies on natural gas, consists of combined- cycle, gas-turbine baseload units and simple-cycle, gas-turbine peaking units (Nexant 2006). Power-sector reform, initiated in 1994, envisions the separation of sector regulation from ownership and operation and independent, commercially oriented sector entities. It calls for the unbundling of generation, transmission, and distribution functions and increased private-sector participation in generation and distribution. Formerly a monopoly operator, the Bangladesh Power Development Board is being converted into a holding company. The Energy Regulatory Commission, established in 2004, is to play a vital role Figure 6.1. Annual Load Shedding, 1993­2004 Source: Nexant (2006). 72 World Bank Working Paper in a range of sector operations, as noted in the following sections, but has been hampered by slow organizational approval processes and lack of funding. Key power-sector entities and their functions are as follows: MPEMR. The Ministry of Power, Energy, and Mineral Resources oversees sector management. ERC. The Energy Regulatory Commission has authority over consumer protection, approval of tariffs and pricing, issuance of generation and distribution licenses, and promotion of competition. Power Cell. Within the MPEMR power division, the Power Cell oversees power- sector reform. BPDB. The Bangladesh Power Development Board operates most publicly owned generators and some urban distributors; it acts as a single buyer, purchasing from public and private generators and selling to distributors. PGCB. The Power Grid Company of Bangladesh, a wholly owned subsidiary of the BPDB, operates the national transmission grid, schedules grid operations, and wheels energy to distributors. DESA. The Dhaka Electricity Supply Authority (not yet functional) distributes energy and conducts commercial operations in Dhaka and adjoining areas, except for Mirpur and Gulshan. DESCO. The Dhaka Electric Supply Company distributes energy and conducts commercial operations in the Mirpur and Gulshan jurisdictions of the Dhaka metropolitan area. REB. The Rural Electrification Board oversees operations of consumer-owned rural electric cooperatives (PBSs). It performs supervisory and regulatory duties to ensure that technical standards are met and performance is monitored. WZPDC. The West Zone Power Distribution Company is responsible for regional distribution in Khulna. IPP. There are some independent power suppliers (IPP) in Bangladesh that sell electricity to the transmission company. EA. Electrical Advisor. CEI. Chief Electric Inspector. Over the next several years, the newly constituted Rural Power Company, which runs a 140-MW generator and sells power to the Rural Electrification Board, plans to own and operate a substantial portion of the power sector's generation assets. As part of the unbundling process, regional distribution companies are being established in Khulna (West Zone Power Distribution Company), Rajshahi (Northwest Distribution Company), Chittagong and Comilla (South Zone Distribution Company), and Mymensingh and Sylhet (Central Distribution Company). Toward a Pro-Poor Approach. The Rural Electrification Board (REB), a semiautonomous government agency, is responsible for electrifying rural Bangladesh. Established in 1977, REB operations are based on the U.S. model of consumer-owned rural electric cooperatives. Today, rural Bangladesh is divided into 70 cooperatives or PBSs (Palli Bidyut Samities), Restoring Balance: Bangladesh's Rural Energy Realities 73 which benefit an estimated 6.78 million households--a remarkable achievement in a country of 140.5 million people. Yet most of the PBSs are not financially viable. Massive operating and capital subsidies are required to keep the program functioning and expanding. Although the PBSs are allowed to set their own tariffs, reflecting local costs, most agricultural and household consumption is highly subsidized. Another key issue is that the rural power sector relies heavily on the Bangladesh Power Development Board for its power supply and the Power Grid Company of Bangladesh for load scheduling. As a result, rural load shedding exceeds urban load shedding, even though the REB network accounts for only 40 percent of national electricity consumption. As this survey's results highlight, providing village access to electricity does not guar- antee access at the household level. The government's Poverty Reduction Strategy Plan has identified various measures to improve the pro-poor orientation of rural electrification. These include: (1) setting up a microcredit mechanism to help poor households surmount the hurdle of first connection, (2) breaking up the REB into regional boards to create a more locally driven approach to the economic development aspects of rural electrification, Figure 6.2. Bangladesh Power Sector Organization Bangladesh Power Sector: Current Structure MPEMR ERC (Ministry of Power, Energy and Mineral Resources) Power Cell EA & CEI Generation BPDB IPP Small IPP PGCB (An enterprise of BPDB) Transmission Distribution BPDB (South/Central DESA DESCO REB/POSs WZPDC REB/POSs /North Zone) Source: Nexant (2006). 74 World Bank Working Paper and (3) increasing PBS financial viability by their taking over peri-urban enclaves still served by the Bangladesh Power Development Board. Oil and Gas Sector Natural gas is by far the most important commercial fuel in Bangladesh, accounting for more than 70 percent of primary commercial energy production. Most oil is imported in the form of refined products. Since 1998, urban marketing and use of bottled gas has grown rapidly. In and around the capital city of Dhaka, bottled compressed natural gas (CNG) is now a cheap substitute for gasoline in the transport sector. Total in-country production of liquefied petroleum gas (LPG) is 23,000 tons per year, with another 35,000 to 45,000 tons imported and locally bottled. Yet only 10 to 15 percent makes its way beyond the peri-urban centers to more remote rural areas. Promoting rural access to natural gas could result in significant welfare gains for many rural households. For example, by switching from biomass to LPG to meet a portion of their cooking needs, rural family members, especially women, could realize large time savings (through fewer hours spent collecting biofuels) and better health (through reduced exposure to indoor pollution). Rural populations' access to pipeline gas is also inadequate. Indeed, rural residents represent only 6 percent of national coverage. By contrast, in major urban areas served by regional distribution companies, non-bulk sale to commercial and residential customers is extensive. One exception to the poor rural distribution network is a US$300 million project financed by the Asian Development Bank that plans to expand transmission infrastructure into the western region surrounding Rajshahi by 2010. Sector Organization. The oil and gas sector is organized into four segments: exploration, production, transmission, and distribution (including marketing) (see Figure 6.3). Figure 6.3. Oil and Gas Sector Organization Source: McKenzie (2006). Restoring Balance: Bangladesh's Rural Energy Realities 75 Key sector institutions and their functions are as follows: Petrobangla. Successor to the Bangladesh Mineral Oil and Gas Corporation, Petrobangla is a public-sector holding corporation responsible for oversight and coordination of 11 subsidiary gas companies. BAPEX. The Bangladesh Petroleum Exploration Company is a state-owned entity that manages exploration and production. BGFCL. The Bangladesh Gas Fields Company Limited is a state-owned entity that manages production. SGFL. The Sylhet Gas Fields Limited is a state-owned entity that manages production. IOCs. The International Oil and Gas Companies are private-sector exploration and production entities; they include Unocal Corporation, Cairn Energy, Niko Resources, and Tullow Oil. GTCL. The Gas Transmission Company Limited manages midstream. TGTDCL. The Titas Gas Transmission and Distribution Company Limited is a state-owned, regional distribution entity. BGSL. The Bakhrabad Gas Systems Limited is a state-owned, regional distribution entity. JGTDCL. The Jalalabad Gas Transmission and Distribution Company Limited is a state-owned, regional distribution entity. PGCL. The Pashchimanchal Gas Company Limited is a state-owned, regional dis- tribution entity. RPGCL. The Rupantarita Prakritik Gas Company Limited is a CNG marketing entity. Downstream Deficiencies and Reform. The downstream gas sector includes pipeline gas, bottled LPG, and CNG. Downstream sector performance is hampered by network losses of about 20 percent, as well as pricing and commercial performance deficiencies. Fiscal drain due to inadequate pricing prevents operating companies from reinvesting in required exploration and production facilities and network expansion.7 Although LPG is widely available in most peri-urban areas, it is estimated that only 10 to 15 percent finds its way to rural areas for consumption. In more remote rural areas, where LPG prices are fixed, it is not unusual for local dealers to charge significantly higher prices. The government has agreed to a roadmap for sector reform, which emphasizes clear separation of policy making, regulatory, and service provisions through the creation of well-structured, adequately funded entities. Key decisions include transferring Petrobangla's quasi-regulatory functions to the newly formed Energy Regulatory Commission and resolv- ing retail-market arrears to alleviate fiscal drain on Petrobangla. Renewable Energy Sector Non-biomass renewable-energy systems also hold promise for rural Bangladesh. Solar photo- voltaics (PV) are used widely throughout the country, with more than 80,000 reported household- and enterprise-level installations (3.5 MW total capacity). Solar PV could be used within a broader rural electrification program if affordable products that meet consumer needs were supplied and supported. 76 World Bank Working Paper Other potential renewable-energy sources include wind, hydropower, geothermal, tidal, and wave. Although wind energy has made some inroads, its potential is likely limited to coastal areas and islands with strong wind regimes; currently, the country has 2 MW of installed wind turbines in operation. The potential for micro- and mini-hydropower is also limited (with the exception of Chittagong and the Chittagong Hill tract); the country has one large facility at Kaptai, installed some 40 years ago, which produces 1,000 GWh per year. Until recently, renewable energy was developed on an ad-hoc basis. Over the past 20 years, various government agencies, research institutes, NGOs, and private-sector companies have participated in development, demonstration, and implementation. Current institutional arrangements reflect the sector's nascent state. Key sector institutions and their focal areas are as follows: Grameen Shakti. A subsidiary of the Grameen Bank, Grameen Shakti focuses on solar home systems (SHS), having installed 5,000 systems over the past two years; household biogas systems; and improved stoves. REB. The Rural Electrification Board centers on household- and village-scale, off-grid PV systems; it installed nearly 2,000 SHS through the PBSs. IDCOL. The Infrastructure Development Company Limited focuses on household- and village-scale, off-grid PV systems. IFRD. The Institute of Fuel Research and Development, within the Bangladesh Council for Scientific and Industrial Research, centers on small- and medium-sized biogas digesters. Local Government Engineering Department. The Department handles small- and medium-sized biogas digesters. BRAC. The Bangladesh Rural Advancement Committee, the largest national NGO, centers on SHS, having supplied nearly 1,200 systems, and biogas installations. Proshika. A leading NGO, Proshika focuses on biogas installations. To date, Grameen Shakti and private-sector companies have led much of the work in renewable energy. In addition, the World Bank has supported the scaling up of solar PV systems. Other donors have included the United Nations Development Program and the U.S. Agency for International Development. Building an Integrated Approach The 1996 National Energy Policy, developed by the Ministry of Power, Energy, and Mineral Resources as part of power-sector reform, represented a policy milestone in visualizing the country's future energy needs. But it failed to comprehensively address rural energy as a unique and important energy-sector issue. Although the policy covered such critical areas as rural electrification, rural penetration of commercial fuels, and reforestation, it overlooked improved use of biomass, biogas production, and other programs equally vital to energizing rural Bangladesh. The 2006 policy revision takes a more comprehensive approach, focusing on integration across subsectors, demand-side issues, more effective institutional arrangements, and policy monitoring. Among the critical rural energy issues identified are the financial viability of Restoring Balance: Bangladesh's Rural Energy Realities 77 loss-making PBSs, indoor air pollution, and the effects of commercialized biomass supplies on the poor. Tackling rural energy issues in a more integrated way calls for establishing a Rural Energy Steering Committee within the Ministry to improve linkages between energy and rural development institutions. The Sustainable Energy Development Authority (SEDA), the proposed secretariat for the Committee, would coordinate with ministries and government agencies to develop and oversee policy and program implementation. It is envisioned that the new Committee and SEDA would work together to recommend ways in which to improve coordination between all levels (national, district, subdistrict, and local) and form local organizations that could plan and implement sustainable energy projects. Summing up, the many diverse institutions currently addressing rural energy issues in Bangladesh are committed to providing rural residents better energy services. But the multitude of institutional players creates a significant role for improving inter-institutional coordination. That this role has often gone unfulfilled may explain why rural energy remains invisible to many policy makers. Suggested steps toward filling this institutional gap are the focus of the final chapter. Notes 1. In 2005, Bangladesh's National Strategy for Accelerated Poverty Alleviation set forth eight strategic priorities, most of which imply a role for promoting better energy policies to improve rural access. 2. At the local level, for example, transportation costs and seasonal irrigation requirements call for a degree of local supply­demand balance or self-reliance. 3. The government's current pricing of electricity, oil, and gas is far below the cost of supply. 4. In deciding whether subsidies are required for particular types of energy or end users, policy makers must consider both the social and economic costs and periodically revisit the results as these factors change. 5. Paradoxically, the rural buying and selling of biomass also creates jobs. Indeed, employment from trading biomass exceeds that from trading substitute fuel products (e.g., kerosene, LPG, and electricity). Thus, in the push to promote modern energy, one must also consider the potential loss to rural employment, which could lead to even greater poverty. 6. Although the daily cow-dung input for cooking three meals a day is known, reports conflict regarding the number of cattle owned by rural households (BBS 2005). Anecdotal evidence suggests seasonal, rather than year-round, ownership. In addition, use of power tillers and tractors is widespread. 7. Petrobangla buys IOC gas at US$2.6­2.9 per MCF but charges domestic customers only US$1.40 per MCF. CHAPTE R 7 Policy Recommendations T his study--the first in 30 years to examine energy access in rural Bangladesh-- reveals several stark realities. Rural residents continue to rely heavily on traditional biomass to supply most of their energy. But biomass used in traditional ways has energy efficiencies of only 15 to 25 percent; thus, extensive time is required for fuel collection and meal preparation. Furthermore, biomass cooking fuels result in indoor air pollution (IAP), which is linked to respiratory disease and related health risks. Rural residents face an increasing fuelwood shortage, meaning that many are relying on residues, including dung, leaves, and grass, to meet their basic cooking needs. Rural businesses also depend on biomass as one of their main energy sources. Thus, the problems associated with biomass energy are widespread. Significant bottlenecks prevent most forms of modern energy--a main alternative to biomass fuels--from reaching many rural people. Although well managed, rural electrifica- tion has reached less than one-third of the population. Households and businesses fortunate enough to have electricity must contend with frequent power outages, which imply added expenditure on backup diesel or kerosene fuel and hindrances to rural productivity and quality of life. That liquefied petroleum gas (LPG) is used little for cooking reflects its high cost and lack of availability. But the rural energy picture is not all bleak. As this study reveals, significant benefits, including more lucrative rural businesses and better educated rural households, can result from promoting access to modern and sustainable energy. Thus, the development of energy infrastructure is sound development policy that policy makers should actively promote. With the exception of rural electrification, Bangladesh's energy policies and institutions have focused mainly on urban markets. In 2006, the country still lacked a comprehensive, unified rural energy policy with appropriate institutional support. As a result, the rural energy issue remains invisible to many policy makers. 79 80 World Bank Working Paper The major recommendations for the rural energy sector in this study focus on devel- oping the capacity for implementing high-quality programs rather than specific policies. Energy policies cannot be static, given the ever-changing energy markets around the world. The main recommendations are as follows: There is a need to further develop the institutional capacity to deal with biomass energy problems, including the strengthening of the Infrastructure Develop- ment Company Limited (IDCOL) and perhaps other agencies to diversify into such areas as biogas, improved stoves, and other household energy technologies. In addition, more attention needs to focus on improving local biomass supply in and around farms. The recent progress in off-grid electrification and renewable energy through the IDCOL program is encouraging and should be continued and diversified. The Rural Electrification Board needs to maintain its distance from political pressures. It also needs to deal with the problem of brownouts and blackouts prevalent in the program and assess ways to financially strengthen the country's weaker rural electric cooperatives or PBSs. Given the spread between village and household electrification and the low number of agricultural pump connections, an assessment of whether financially viable electricity intensification policies might be pursued is needed. Given the extensive commercialization and increasing scarcity of quality biomass energy in rural areas, there appears to be a market for liquefied petroleum gas (LPG), with the right marketing strategies. Currently, the reach of LPG into rural areas is low compared to other South Asian countries. Finally, given the government's weak capacity to assess rural energy policies in an objective way, it is recommended that a policy group be created via an independent agency, part of another agency, or a consortium of existing research institutions. It is also recommended that a more complete set of energy questions be incorporated into the national income and expenditure surveys, as purchase and collection of biomass energy are nearly absent entirely from existing ones. The sections that follow bring the broad array of fuels that sustain rural people's lives into focus. Tackling the constellation of issues surrounding their effective use is critical to the sustainability of Bangladesh's economic future. But the task will not be easy. It will require complementary efforts on many diverse fronts. Problems and Potential of Biomass Energy Biomass--a vital energy source in rural Bangladesh--is often overlooked by energy policy makers. As this study shows, biomass is the country's primary rural household fuel. More than 95 percent use it to cook all or part of their meals. Moreover, most household- and village-based enterprises use biomass for heating. As the survey findings illustrate, beneath this biomass dependence, an energy crisis is simmering. In regions experiencing a continuing decline in the availability of quality biomass, many people seem to be turning from fuelwood to dung, straw, tree leaves, and grass. As the previous chapter discussions underscore, quality biomass in the form of wood Restoring Balance: Bangladesh's Rural Energy Realities 81 is fast becoming monetized, a reflection of its scarcity. In areas where wood is not purchased, all family members, but especially women, spend significant amounts of time (nearly 200 hours annually) collecting all forms of biomass. Such arduous chores divert time from competing tasks and important human-development activities. Moreover, poorly designed biomass-burning stoves without chimneys have low energy efficiency. Used indoors, such stoves emit high levels of smoke and other products of incomplete combustion that collect at high levels indoors. The result is significant health problems for all family members, especially women and children, who spend many hours indoors within the vicinity of the stove. Mitigating the ill effects of biomass burning requires both demand- and supply-side solutions. Because all biomass shortages vary by region, the solutions must be local; thus, one should caution against any blanket interventions covering the entire country. Surveys during various times of the year might be undertaken that ask household members a range of questions regarding the availability of biomass (both collected and purchased), alternative fuels, cooking devices, food preparation methods, and kitchen practices. Various proposed solutions could then be offered householders, local leaders, and non-governmental organi- zations (NGOs). The participation of women would be especially important as they are the ones who deal daily with biomass collection, processing, and cooking. With these groups' active participation and cooperation, the most suitable solutions should be found. Biomass demand can be reduced by using improved cookstoves and kitchen practices. Where possible, fuel substitution including the use of kerosene or LPG should be encour- aged. Biomass supply can be increased by planting trees around homes, maintaining and improving local natural resources, and increasing agricultural productivity. Although the impetus for improvement should come from rural residents, the government needs to play a supportive role through such activities as stove testing and quality control, providing seedlings and other inputs, and offering extension advice and demonstrations of relevant technologies. Because biomass is an important fuel in rural industry and service sectors, improving intermediate and end-use efficiency in these sectors is essential. The increased commer- cialization of biomass fuels is a double-sided issue: In areas where commercialization has accelerated, many rural residents face shortages of locally available, quality fuelwood; conversely, a certain proportion earns income through biomass growing, production, transport, and trade, which should assist in alleviating poverty.1 In short, promoting a sustainable supply-demand balance of biomass energy, supported by appropriate policies, should be encouraged. To this end, rural residents can be encouraged to move toward more efficient use of biomass energy, in combination with tree planting on farmsteads and improved local forest management. Increasing Biomass Efficiency and Use This section reviews past and ongoing initiatives that have aimed to improve biomass energy efficiency and use in rural areas, highlighting lessons and potential solutions that can mitigate risks to human and environmental health. Improved Stoves: Potential for Revitalization. Some 20 million Bangladeshi households own up to 40 million biomass stoves, most of which are inefficient and produce excessive 82 World Bank Working Paper smoke and other products of incomplete combustion. As this study shows, millions of people in close proximity to such stoves suffer adverse health effects. At the same time, the products of incomplete combustion contribute to the accumulation of greenhouse gases. For these reasons, all relevant ministries--health, environment, energy, and forestry-- should be much concerned about the human and environmental damage caused by using biomass energy in inefficient, polluting ways. Worldwide, 1.5 million people die prematurely each year because of indoor air pollution (IAP), caused by using inefficient biomass stoves (WHO 2006). Many millions of days are lost through sickness caused by inhaling excessive smoke. Although biomass drying, improved kitchen practices, and better ventilation can mitigate the negative health effects, such measures cannot reduce biomass demand. More sustainable solutions call for efforts that increase stove efficiency and vent smoke from people's living areas.2 There are some recent estimates in Bangladesh that IAP might account for as much as 8 percent of the burden of disease. In Bangladesh, most intervention programs to popularize improved stoves were proven ineffective or were small in scope. The largest program, terminated in 2001, was that of the government of Bangladesh. The program was supported by various organizations, including the Institute of Fuel Research and Development, the Bangladesh Council for Scientific and Industrial Research, Ansar-VDP, and the Bangladesh Rural Development Board. By December 2002, an estimated 300,000 Bangladeshi households had received access to improved stoves. Based on these penetration rates, one could expect that about half a million households would have gained access by 2007. This figure is less than 0.5 percent of the population; thus, there is much room for improvement (Hossain 2003). More recently, various agencies have continued to promote improved stoves, albeit on a smaller scale. Current NGO initiatives are not widespread for lack of trained personnel and funds. Grameen Shakti has been working throughout the country on initiatives funded by international donors, whereby beneficiaries are provided microfinance on demand; to date, its accomplishments have been modest. The Village Education Resource Center, another NGO striving to alleviate poverty in Bangladesh, is among the few large ones actively working on improved stove initiatives. The Center operates from its head office in Savar and 14 branch offices located across various districts. Its main focal areas are water, sanitation, hygiene, and improved stoves. It has pioneered a participatory assessment method to mobilize communities and generate village demand for improved stoves, house modifications, and behavioral change to transform villages into smoke-free ones. The Center's demand-oriented approaches and stoves' designs cater to consumer requirements. Even so, accomplishments to date have been modest. Finally, the Local Government Engineering Department, the technical arm of the Local Government Ministry for Rural Infrastructure Development with nationwide reach, is implementing work on improved stoves. Various reasons account for the ineffectiveness of many earlier government initiatives. In some cases, the stoves were inappropriate to the needs of the cook. In others, programs were driven by installation targets with little follow-up activities. Many programs were subsidized, with benefits going to the organizers or installers rather than the recipients. In still other cases, the improved stoves were no better than the originals, but were less versatile. Where improved stove programs have succeeded, cooks have been an integral part of the process. Advice can be offered about the pros and cons of interventions. Saving time and money are usually top priorities; at the same time, comparing the health effects of various stove designs and kitchen layouts should be stressed. Improved stoves need not be Restoring Balance: Bangladesh's Rural Energy Realities 83 complicated; they can be made from local materials. Local women's groups can be organized and trained to make and install improved stoves on a commercial basis after deciding on household cooking needs. Such groups can fabricate chimney pieces (if not purchased) from mud, using straw as a reinforcing agent and burnt clay inserts of standard sizes, thereby reducing the cost US$5­10 per stove. Evaluation research suggests that better-coordinated efforts could revitalize the potential of improved stoves, which would reduce pressure on traditional biomass supplies and significantly improve the health of biomass-reliant rural populations. It is suggested that improved stove programs take a more pragmatic approach with regard to stove benefits and performance, including demonstrations and promotion to increase awareness of the dangers of IAP and need to conserve biomass fuels. Undertaking a sustained nationwide program requires sufficient administrative and technical infrastructure. Specific activities include the following: stove design and testing; demonstration; research and development; quality control; promotion in all media forms; making simple tools available to stove makers; providing advice on material sources, use, and suitability (especially if clay); training of stove makers, repair-service operators, and extension workers; monitoring and evaluation; and loans. End-user feedback on the pros and cons of various stove designs should be ongoing, and suggestions should be integrated into the further refinement of stove design. Stove initiatives should be part of a larger program to improve village life, especially the health and welfare of community members. Part of this program should provide train- ing on biomass use (e.g., drying and using in appropriate sizes and discussion on food preparation and energy-saving methods). Many such initiatives could be initiated within the community; through these efforts, the entire village could participate and take pride in its achievements. The overall program should promote market-based approaches that rely on local pro- ducers and distributors who profit only from satisfied customers. Direct stove subsidies should be avoided, as they create market distortions and ultimately hinder stove commer- cialization. Instead, subsidies should be directed toward activities that support the program (e.g., greater government backstopping, active NGOs, small-scale enterprises, and using women's groups to promote improved stoves for rural households and microenterprises). Finally, a wide range of stove designs (some of which would be suitable for the service sector [e.g., canteens, restaurants, and hot-food shops]) is needed. Biogas Digesters for Cooking and Farm Productivity. Currently, some 25,000 biogas digesters are operational in Bangladesh. Each digester requires dung from about six cows (i.e., assuming that each produces 10 kg of wet dung per day, equivalent to 1.15 kg of air-dry dung). If 25 percent of cow dung could be utilized, approximately 1 million family-sized digesters, each supplying cooking fuel for five to six people, could be built. Unfortunately, many farmers lack access to the required amount of dung, which hinders large-scale introduction. Passing dung through digesters has several advantages. First, the gas produced is easy to light and control; its efficiency for cooking is 50 to 60 percent, depending on the stove. Used for lighting, the gas emits a brighter light than kerosene. Most pathogens in the dung are destroyed during the production process, and the resulting slurry makes an excellent fertilizer. If dung is applied directly to paddy fields, methane, a potent greenhouse gas with 84 World Bank Working Paper 20 times the warming effect of carbon dioxide (CO2), is produced and vented into the atmosphere. Digesters capture this methane and use it for cooking; furthermore, slurry is a better fertilizer than dung. The Bangladesh Council for Scientific and Industrial Research (BCSIR) runs the government's active but modest biogas digester program, which focuses on relatively large units for which farmers receive advice and a subsidy of Tk 7,500 (US$120) per digester. Because these digesters are relatively expensive to build, they must be properly maintained to perform effectively. A fixed-dome, 5 m3 unit costs US$240­290. Cheaper black plastic (PVC) models that fit into trenches are variously sized according to the number of cattle. The cost of such digesters (with appliances) is about US$100. Farmers with only one or two cows could use this type of digester to meet a portion of their cooking needs and, at the same time, produce an excellent fertilizer. Financing could be arranged through the active program of a Bangladeshi microenterprise. Farmers or villagers could cooperate to build community digesters, which could provide enough fuel to run electric generators for part of each day. Over the next several years, the Infrastructure Development Company Limited (IDCOL) plans to facilitate the construction of thousands of biogas plants. It might be beneficial for two complementary agencies, such as the BCSIR and IDCOL, to foster biogas development. Higher Bioenergy Efficiency for Industry. In 2000, the formal and informal industrial sector--mostly agro-processing industries--accounted for more than 20 percent of total biomass energy demand. Energy uses included parboiling of rice, sugar production, tea drying, tobacco curing, and baking. Non agro-based industries included brick and tile manufacturing, lime burning, road tarring, and soap making. Wood was the main input for charcoal production. Most industries were based in rural areas, relying on inexpensive, readily available biomass supply, characterized by low fuel efficiency. It is recommended that the government provide a range of technical assistance, including energy audits, to improve fuel efficiency. In addition to improved stove initiatives, interventions should be pursued for charcoal production and other activities where solid biomass is used or produced. Technical assistance should focus on a wide array of areas: improving institutional- and service-sector stoves; improving brick, pottery, and lime kilns; conducting energy audits at sugar factories and sawmills to assess energy-saving approaches; listing agro-processing factories that use biomass energy (e.g., tea drying, rice parboiling, and fish smoking); and examining ways to improve energy efficiency for all biomass users. To succeed, these initiatives require quick payback periods. Charcoal producers could be trained to better manage wood resources and improve production processes. If forests are the main wood source, the forest service should work with charcoalers to plan a cutting cycle. Cooperating with producers would enhance forest sustainability and diminish illegal cutting. With regard to briquetting of difficult-to-burn residue (e.g., husks and sawdust), many past initiatives have failed because of inappropriate, sophisticated equipment or lack of sound marketing and management. The cost of briquetting is generally expensive, and the product may not be competitive with unprocessed biomass. Hand machines and locally made, inexpensive binders, as well as potential markets for the finished product, should be investigated. Densification may offer a solution for straw, which is bulky and quick-burning, but price is a constraint. Various types of densifying equipment, including hand presses, as well as binders, should be compared. If unemployment and underemployment are problems, Restoring Balance: Bangladesh's Rural Energy Realities 85 simple hand devices may prove more appropriate than high-tech machines. Recently, this area has become more widespread, with several hundred factories operating across the country. Increasing Biomass Supply A complementary solution to Bangladesh's biomass energy problem is to increase or maintain supply through local forest management or growing of trees, bamboo, and shrubs on and around farms and along roads. Trees grown outside the forest are a principal source of rural fuelwood. Such trees are usually planted close to the demand center and are intensively managed, especially if privately owned. They provide added income by yielding multiple products (e.g., fruits, medicines and herbs, and nectar and leaves used as feedstock in honey and silk production). The upsurge in tree-planting activities over the past 20 years has resulted from the support of NGOs, government, and international donors, as well as the active involvement of rural people. Through participatory silviculture, it will be possible to increase tree planting and management both in and outside the forest. It comes as no surprise that agricultural conversion threatens a portion of rural Bangladesh's forests and grasslands. Rural population growth is 2.5 million per year, more than 62 percent of the country's annual growth rate of 4 million. Unless agricultural productivity keeps pace with population growth, forested lands will be cleared for food production. Short-rotation, nitrogen-fixing trees could be planted in and around farmland to maintain, if not increase, productivity and provide farm animals browsing material. Such trees yield stick wood; and increased productivity and browsing may result in more agricultural residue and dung. Thus, it is essential that the relevant ministries, particularly agriculture and forestry, work jointly to facilitate increased agricultural productivity, without which the biomass resource base will surely diminish. Biomass is rural Bangladesh's most important cooking, heating, and industrial fuel and will remain so well into the future. For perennials, this implies management that removes not more than annual growth over the long term; for annuals, it means, at a minimum, that soil fertility be maintained so that land does not deteriorate to a state of marginal or no productivity. The expected doubling of Bangladesh's population over the next 30 years will further strain the natural resource base. Adopting an effective strategy is urgently needed, especially in rural areas, to ensure an ongoing increasing supply of renewable biomass energy. Imple- mentation of this strategy should be guided by powerful institutional voices within the government. Thus, in addition to the Ministry of Power, Energy, and Mineral Resources, the ministries of agriculture, forestry, health, rural development, women's affairs, and others are also essential. Especially important is the proposed Renewable Energy Development Agency (REDA), backed by the Ministry's renewable energy division. Proposed Strategy Given Bangladesh's significant dependence on biomass energy, we reiterate some of the previously cited suggestions to conclude this section. A strategy to improve biomass energy prospects in rural areas could take various approaches, four of which are highlighted here. First, the intermediate and end-use efficiency of biomass energy production should be 86 World Bank Working Paper increased, especially at the household level. The aim would be not only to reduce unit energy consumption, but, more importantly, to decrease products of incomplete combustion, which damage human health and contribute to the accumulation of greenhouse gases. Second, biomass production (of both annuals and perennials) should be increased on all land formations so that the country can keep pace with demand for biomass and its products. Ideally, increased productivity should outpace population growth so that more biomass can be used for all purposes, especially energy. Third, use of modern energy and more convenient forms of biomass should be promoted, despite ongoing use of traditional forms of biomass, which will continue for many years to come. (Modern forms of energy include solid, liquid, and gas products used as fossil-fuel substitutes; biomass and its products could be used to generate electricity, where appropriate). Finally, buildings and habitats should be designed to conserve energy. Trees planted around buildings and along streets would reduce the ambient temperature, which would, in turn, reduce the need for cooling in offices and homes while periodically providing biomass. Trees would also mitigate erosion and improve the environment. Within the country's current rural energy policy and institutional framework, these recommendations may be difficult to implement. Thus, the last section of this chapter turns to the complexity of institutional issues that must be tackled in order for policy implementation to succeed. However, we now turn to a successful program to promote electricity in Bangladesh. Rural Electrification and Rural Development In 1976, Bangladesh was a newly independent country grappling with the challenge of creating national policies and programs amid an underdeveloped infrastructure, rapidly growing population, and frequent natural disasters. Few would have predicted that 25 years later, this poor South Asian country would have succeeded in providing electricity to nearly 30 percent of its rural households. Bangladesh's story of rural grid electrification is important for several reasons. First, despite relatively low coverage rates compared to other developing countries, an additional 600,000 rural customers receive electricity services each year. This annual increase exceeds the total rural population of many countries. Second, having a well-managed, well-run pro- gram is somewhat unique in South Asia. The rural cooperatives have low distribution system losses of only 15 to 17 percent; their 95 percent revenue-collection rate is high by developing country standards. Third, though one of the world's poorest countries, Bangladesh demon- strates how rural electrification programs can succeed under adverse economic conditions. A complementary rural electrification story is Bangladesh's recent effort to strengthen capacity in decentralized or off-grid service. Although the grid is expanding service by large numbers every year, systems expand outward from generation stations, and transmission lines serve the most highly populated areas first. As a result, some 50 percent of rural villages--70 percent of rural people--remain without a connection. But a host of renewable technologies now available for decentralized generation and provision offer promise that many households that otherwise would remain unconnected for years will be served. This section explores ways in which to improve both grid and off-grid electrification programs. Restoring Balance: Bangladesh's Rural Energy Realities 87 Grid and Off-grid Programs Both grid and off-grid electrification programs are important to the socioeconomic devel- opment of rural Bangladesh. The evidence is convincing that electrification translates into substantial gains in household welfare and a higher quality of life. Indeed, the benefits of lighting alone are highly valued by rural households that adopt electricity. As this study shows, household- and village-based microenterprises that connect to the grid are more profitable than comparable businesses served by private or locally purchased generators. Cooperative Grid Electrification: Role of the Rural Electrification Board. One of Bangladesh's most successful development programs over the past 20 years has been the extension of grid electricity in rural areas. As discussed in previous chapters, the Rural Electrification Board (REB) promotes the development of electricity distribution companies based on the rural cooperative model adopted in the United States. In 1977, it was argued that enlarging the public-sector electricity company's focus to include low-density, poorer rural areas was unfeasible, given the company's already overextended efforts to serve urban areas. Thus, the REB was created in 1978 as a semi-autonomous public agency to promote electricity in rural areas and implement a financially viable program. As an agency of the Ministry of Power, Energy, and Mineral Resources, the REB is responsible for planning and implementing all investments in rural electrification infrastructure, overseeing performance of the rural electric cooperatives or PBSs, and regulating prices. Over its 25-year history, the REB has constructed the infrastructure for Bangladesh's entire national rural electrification system. In this capacity, a major respon- sibility has been to manage loans and grants provided by international donors with which to finance an infrastructure development program. Another essential function has been to finance short-term capital needs of newly established PBSs. Finally, the board has provided the PBSs vital technical assistance and monitoring of technical and financial performance, thus helping them evolve into modern distribution utilities. Despite the REB's long success in promoting rural electrification, recent problems call for urgent attention. As this study's national survey reveals, rural consumers perceive that electricity supply has become unreliable. Some 80 percent of rural households report daily outages, while 60 percent report significant power fluctuations. Survey results also bring into question the REB's selected coverage approach, as 20 percent of households that receive electricity indirectly from the PBSs--most commonly via a neighbor--are billed at a higher rate than that charged by the PBSs. While not illegal in the strictest sense (i.e., electricity is metered and paid for by the collecting household), this practice suggests that some type of barrier prohibits households from receiving electricity directly from the PBSs. It may be that households are beyond the required distance from the grid or informally connected ones want to avoid connection charges. The national survey also reveals a significant potential to replace diesel-powered irrigation motors with electric ones. Investment in infrastructure to deliver electricity to rural areas could be complemented by development of daytime loads. Since most current rural demand is during evening hours, revenue from the additional agricultural load based on availability of groundwater potential could significantly improve PBS finances. Recent financial analysis of the PBSs has brought into question the REB's pricing and subsidy policies. In 2006, the PBS service territories differed markedly; some were 88 World Bank Working Paper characterized by lucrative households, while many others served poorer regions. The REB's standard subsidy and pricing policies mean that the financial viability of PBSs in poorer service territories tends to be poor. Amid the constant struggle to meet financial targets, these cooperatives lack the financial resources to serve their customers properly. Thus, careful review of the policies of uniform subsidies and similar tariffs for all PBSs is recommended. One potential solution might include mechanisms that give preferential treatment to companies serving poorer regions, such as permitting higher tariffs or providing electricity at lower bulk prices. The conclusion is that Bangladesh's rural electrification program has been both active and effective in its positive development outcomes. The national survey on rural electrification confirms electricity's significant effect on rural development. In 2006, the PBSs were generally well managed; yet the program has a nationwide generation problem. Rural customers have reported significant outages, which, if they persist, may lead to dissatisfaction with PBS performance. In addition, many cooperatives in more remote areas are not yet financially viable. Policy solutions will require load promotion to increase revenue or price restructuring and subsidies. A recent initiative is studying REB management practices and the board's potential as a freestanding corporation. Such studies stress the REB's key role in supervising and promoting rural electrification for Bangladesh and its need to remain free from political manipulation. Off-grid Electrification: Role of the Infrastructure Development Company Limited. Only 3 percent of Bangladesh's nearly 17 million rural households are gaining access to the national grid each year, creating an enormous potential for off-grid systems. Opportunity is greatest in locales where grid extension is uneconomical. Many rural residents can afford electricity if the costs are spread out over many years. Given that many components of renewable energy systems last 20 years or more, it makes sense to allow households to spread out purchases of renewable energy equipment over a period of years. Such an arrangement could be made in conjunction with rural development programs aimed at improving local incomes. In 2002, the Infrastructure Development Company Limited (IDCOL) was made the country's focal agency for coordinating the off-grid program, the first phase of which has focused on promoting solar photovoltaic (PV) systems. To date, most work has been done by NGOs specialized in microfinance and microenterprise development. The program has succeeded beyond initial expectations. Working with 16 partner organizations, including Grameen Shakti and other NGOs involved in microcredit and solar home systems (SHS), the program has installed more than 80,000 systems over a three-year period. The program has operated on the premise that partner organizations, who have gained the trust of rural residents, are the most efficient SHS delivery agents. Their collection history has proven strong enough to develop a credit line. IDCOL's supportive role has focused on developing consumer awareness of SHS and their potential for rural lighting. Additional responsibilities have included registering participating organizations eligible for assistance, establishing standards, and refinancing up to 80 percent of partner organi- zations' customer loans. This innovative program has demonstrated that, with careful planning and adequate support in the initial stages, partner organizations can gain proficiency in equipment supplier dealings and after-sale customer support. In addition, the significant portion of Restoring Balance: Bangladesh's Rural Energy Realities 89 income that poor rural households have been willing to allocate for basic lighting has underscored the high value they place on the service. Initially, the program aimed to remain technology neutral. Indeed, IDCOL was selected as the implementing agency because it did not promote a single technology. Although SHS has composed most of the program's lending demand to date, a tremendous unmet need for electricity remains. Recently, several other types of off-grid systems have been implemented, but financing has been limited. As a result, a significant need remains to diversify product lines to microgrids, which could be connected directly to the national grid system once the PBSs reach these remote communities. Toward Policies for Household Petroleum Fuels Before independence in 1971, all of Bangladesh's petroleum products were imported and marketed by private-sector companies. Although these companies competed, they were heavily taxed, which translated into higher fuel prices. After independence and national- ization of the petroleum sector, public-sector companies were created to import, refine, and market kerosene and other petroleum products. For the past 30 years, the Bangladesh Petroleum Corporation and its subsidiaries have controlled most aspects of petroleum supply, including its pricing system. The following sections suggest pricing policies for specific fuels, where appropriate, to promote more equitable rural use. Kerosene Bangladesh's kerosene market consists of government suppliers, known as oil-marketed companies, and private retailers. Theoretically, kerosene is priced uniformly across the country, but transport-cost adjustments are made for market distances greater than 40 kilometers from a supply depot. Price increases occur at each distribution stage: from dealer to subdealer to retailer. The more remote the destination, the higher the fuel price. But the price difference between rural and urban areas is not great, and the system for containing kerosene within a price range works well throughout the country. Because most rural households use the fuel for lighting, it is accessible to most at world-market rates; in addition, it is available in quantities sufficient for cooking. Thus, kerosene pricing policies require no substantial revision. Liquefied Petroleum Gas In 1978, the Bangladesh Petroleum Corporation began commercial production and market- ing of LPG. Over the next 20 years, use of LPG grew slowly. In the late 1990s, the government began to allow private companies to import and market the fuel. Today, Bangladesh has three public and six private LPG marketing companies. Private companies, which must pay import taxes and charges associated with the fuel, compete with public companies, which are exempt from such taxes and fees. Despite this price disadvantage, the private companies are viable and are expanding from large urban centers into smaller towns. Notwithstanding such progress, LPG, which is readily available in urban markets, has largely failed to reach rural households. Because LPG is a clean fuel, it could contribute to 90 World Bank Working Paper alleviating IAP. By having the fuel on tap, rural people could reduce the amount of time spent collecting biomass. The difficulties of promoting LPG in rural areas are well documented. Lower-income rural residents with little cash on hand usually cannot afford the high upfront costs associated with LPG. They are unlikely to pay cash for the typical 12.5 kg cylinder or double-burner stove (costing Tk 2,500­3,000). In addition, LPG transport and storage costs (unlike kerosene's narrower range of prices) are prohibitive for many. However, in other South Asian countries, LPG is penetrating rural areas. In India, for example, more than 20 percent of rural households use LPG to meet at least part of their cooking needs; thus, increased availability appears to lead to greater use. Most such house- holds do not switch completely to LPG; rather, they use the fuel for quick-heating items, such as tea. One should note that the poorest rural Bangladeshi households will not adopt LPG extensively; but more prosperous rural areas have greater scope for promotion. Increased use of LPG in rural areas could be encouraged in several ways. First, com- mercial food enterprises could be targeted initially; they currently purchase biomass for food preparation, which requires significant cash outlays. Second, as in many other countries, the upfront cost of LPG stoves and cylinders, which are necessary for fuel use, could be partially subsidized or paid for in installments; lower upfront costs would put stoves within reach of households who otherwise could not afford LPG. Third, large-capacity cylinders, which target mainly urban consumers, could be complemented by a variety of smaller cylinder sizes. Promoting smaller cylinders in rural areas would reduce the cash outlay for refills. All three recommendations would require well-designed strategies in the safe pro- motion of LPG to new markets. Because private companies must keep their prices competitive with those of public companies, they have a disincentive to service rural areas. As a result, they tend to target relatively better-off markets not served by the main public companies. Thus, it would be sound policy to promote a level playing field for all companies marketing LPG. In this way, it is hoped they would be willing to innovate and expand their markets into even more remote areas not currently served. Natural Gas (Methane) Piped natural gas from fields in the Bay of Bengal is used mainly by householders in Dhaka and other large towns. Indeed, the government has reserved natural gas for household use. Under current pricing policy, households are charged a (subsidized) flat monthly rate, irrespective of the amount of gas consumed. Such a policy leads to abuse and waste and discriminates against households without connections in smaller towns and rural areas. It is recommended that the government reverse the flat-rate tariff so that households are charged per unit of consumption. Also, subsidies, which invariably benefit wealthier households, should be removed. The savings from such changes could be redirected to rural areas to encourage rural electrification, increase biomass availability, and foster its more efficient use. At the same time, the brick-making industry has requested a pipeline connection to enable year-round brick and tile production, which would reduce dependence on imported Indian coal and fuelwood (officially banned from use in brick making).3 Although the imported coal varies in quality, its ash content is high, and thus its energy value is low Restoring Balance: Bangladesh's Rural Energy Realities 91 (not much higher than that of dried wood). Because the price of wood is lower than that of coal, wood is used as a starter and booster fuel. But controlling the temperature of the brick stacks is difficult, and the quality of baked bricks is not uniform. The kilns are usually shut down during the rainy season due to fuel shortages caused by flooding in low-lying areas. If the kilns were fired by natural gas instead of coal, they could operate year-round. In addition, they could be designed to provide uniform heat, ensuring a superior product. Bangladesh's brick manufacturing association has expressed a willingness to pay a non-subsidized price for natural gas, which would generate more government revenue, save foreign exchange, and reduce fuelwood use. It would also mean that a percentage of brick makers could install more-efficient kilns and produce more uniform bricks and tiles. It is recommended that the government examine the merits of supplying brick makers natural gas. Enhancing Rural Energy Projects, Policies, and Strategies Rural energy is a complex issue, encompassing a broad and diverse spectrum of resources-- from petroleum fuels and coal to biomass and renewable energy--spanning multiple sectors, including forestry, electricity, and health. To date, many of Bangladesh's institutions involved in rural energy have not been coordinated. One major recommendation resulting from this study is to develop the policy capacity to tackle rural energy issues in all their complexity. This policy capacity could be used to objectively analyze and promote rural energy solutions and would complement the existing institutional programs that are being used to implement rural energy in Bangladesh. Donors have often advocated projects with a narrow technology focus. But experience teaches that focusing on single technologies does nothing to further markets and private- sector companies to support rural energy development. Such an ad-hoc arrangement is not conducive to capacity-building. Whatever experience is gained under a specific project cannot be applied to subsequent ones for lack of continuity. Moreover, single efforts generally seek exceptions to regulatory policies but are unable to change them. As a result, it is not possible to see the long-term effects of such projects in the form of greater access to quality, rural energy services. Thus, a rural energy institution could provide government advice on how to better focus use of donor funds. To date, successful rural energy programs in Bangladesh have combined financing, institutional support, local support, appropriate pricing, competent implementing firms, and market development. They also have been characterized by a strong coordinating agency. A case in point is the REB; given that it has succeeded in moving the country's electrification rate from 10 to 30 percent, continued support of its efforts to fulfill its ambitious mandate is critical. Inevitably, such institutions face serious challenges as their programs mature. It is imperative that international agencies stay involved to support them through the critical transitions. As noted in previous chapters, the REB now faces an array of complex issues, including differential subsidies for cooperatives that service poor-load areas, takeover of towns and small cities within the boundaries of cooperatives, and the transition to greater self-reliance. Another major challenge identified by this study is access to reliable supplies of bulk electricity. 92 World Bank Working Paper Beyond the REB, institutions that support rural energy have been working relatively independent of one another. IDCOL's recent success in promoting solar PV systems suggests that renewable and other areas of rural energy require a level of support on par with that of grid electrification. Unlike the REB model, that of IDCOL is more decentralized, and there is no overlap between the two. Both models handle financing, provision of technical assistance, lobbying on regulatory issues, vendor approval, and support of venture devel- opment. Neither the REB nor IDCOL limits its focus to technology dissemination. Both institutions are concerned with development of the market and supporting environment for delivery of rural energy services. The recent move toward establishing REDA, while a welcome development, perhaps, does not go far enough. The recommendation is to designate an agency to handle the entire range of rural energy issues, sometimes acting as an executing agency and, at other times, playing a coordination role. Given IDCOL's solid track record, a first step might be to strengthen its off-grid electrification unit, thereby enabling it to handle a broader range of rural energy issues. Because Bangladesh's entire range of rural energy responsibilities is divided among a number of agencies or institutions, development of an active policy unit is needed to advise the government on the diversity of rural energy problems. Such a policy unit, which could be called the rural energy policy unit, would be responsible for formulating and promoting rural energy policies, ranging from biomass to electricity and LPG. It would work hand in hand with various agencies with major responsibility for energy in rural areas, including the REB, IDCOL, and others. There is a significant need for a group in Bangladesh dedicated to evaluating and pro- moting rural energy policies. It is envisioned that Bangladesh's rural energy policy unit would eventually be independent or semi-independent of the government. Over the short term, however, it might be practical to house the unit within an existing agency or public research group. Its role would be to promote sound policies for rural areas and innovative pilot projects; it would not be involved in project development, which would be left to other agencies. In addition to providing policy support to promote rural energy, this new agency could compile and publish rural energy information. Analysis of data from the national survey undertaken for this study--the first of its kind in 25 years--would provide future market opportunities by identifying priority locations for project financing. This would be a great service to all cooperating organizations interested in rural energy, including the many above-mentioned ministries and agencies. It would be imperative to ensure that the tech- nologies recommended were of sufficient quality to meet rural energy needs. In addition, the program scope would need to extend beyond renewable energy to include conventional off-grid power systems in remote areas. Technology coverage should be broadened to include renewable energy (e.g., micro-hydropower, PV, and wind), improved stoves, tree planting, and small diesel systems. Summing up, rural energy is a complex and, for the most part, unprofitable business. But by using a combination of loans and subsidies, both large- and small-scale businesses can become viable to better promote a wide range of rural energy services, from LPG and grid electricity to improved stoves and tree growing. Moving forward requires a combination of research, production and delivery, support, and monitoring and evaluation, all of which must be done through a variety of businesses--including rural electric cooperatives, NGOs, Restoring Balance: Bangladesh's Rural Energy Realities 93 private-sector companies, and local municipalities--interested in serving rural energy customers. The Way Forward Rural energy's importance to the Bangladesh economy cannot be underestimated, given the world's focus on globalization and market reform. This study underscores how the effects of rural energy cut across multiple, diverse facets of rural life--from income and labor productivity to education and women's health. The problems rural people face in obtain- ing safe, clean, and reliable energy supplies are not minor inconveniences. On the contrary, they represent a significant barrier to rural economic development and improved social well being. A multifaceted approach to solving Bangladesh's rural energy problems is not only warranted; it is an essential building block to propel the country into the twenty-first century. The past two decades have witnessed many attempts to promote rural energy. Donor- and public-sector supported projects that have introduced and popularized improved biomass stoves have yielded only limited success, despite the large potential benefits of sustaining biomass supply and improving human health. Social afforestation programs initiated over the period have run their course. Although biogas programs have enjoyed considerable success, they fall far short of realizing their considerable potential. Renewable energy efforts, especially the popularization of solar PV, have achieved a remarkable measure of success; even so, the technical and socioeconomic issues associated with scaling up household and village electrification require capacity-building at national and local levels. The rural energy issue identified is this: Bangladesh has a comprehensive need for better institutional coordination and attainment of a critical mass of technology and market development. With effective institutional coordination, combined with market development, appropriate subsidy and pricing policies, and government and donor support, current and proposed programs can succeed beyond expectations. The call for action is urgent, not only for rural development, but for the country's equitable economic growth. Notes 1. This study examined the purchase of biomass energy, but did not investigate the market structure of fuelwood or other biofuels; thus, more research in this area is needed. 2. It should be cautioned that even the most effective improved stoves being sold or distributed in developing countries do not lower IAP levels to those recommended by international environmental agencies; but it is a step in the right direction. In developed countries, there are stoves for heating that eliminate virtually all indoor pollution and vent little pollution outside the home. 3. Bricks and tiles require less energy than concrete and corrugated iron roofing; thus, brick making should be encouraged. ANNEXES ANNEX 1 Selected Tables from the Household Survey Table A1.1. Survey Sample (number) Division Survey Type Dhaka Rajshahi Chittagong Khulna All Household 603 600 640 548 2,391 Home enterprise 50 27 31 29 137 Village enterprise 85 58 63 136 342 Institutions 38 32 36 50 156 Source: BIDS Survey (2004). 97 98 World Bank Working Paper Table A1.2. Household Distribution by Ownership of Agricultural Assets (percent households) Division Asset Type Dhaka Rajshahi Chittagong Khulna All Land (acres) <0.50 57.9 59.5 56.1 48.7 55.7 0.50­2.49 28.0 27.5 27.0 32.9 28.7 2.50­4.99 8.6 7.0 8.4 12.4 9.0 5.00 5.5 6.0 8.4 6.0 6.5 Non-land Business enterprise/shop 9.8 5.3 6.3 8.6 7.4 Bicycle/motorcycle 15.1 30.7 13.1 28.1 21.5 Rickshaw/van 8.1 9.0 5.0 8.4 7.6 Push cart /bullock cart 1.8 1.0 2.2 4.2 2.3 Boat/engine boat 12.9 6.5 2.5 8.0 7.4 Irrigation pump 4.6 13.5 3.0 4.7 6.4 Tiller/tractor 1.3 3.2 0.0 1.3 1.4 Thresher 0.5 1.5 3.0 3.8 2.2 Rice/flour mill/cane crusher 1.2 0.5 0.0 1.6 0.8 Hand tube well 45.6 60.2 30.5 17.5 38.8 Other 7.6 4.3 17.3 10.9 9.7 None 32.7 24.5 42.5 42.2 35.4 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 99 Table A1.3. Household Distribution by Various Characteristics (percent distribution) Division Household Characteristic Dhaka Rajshahi Chittagong Khulna All Education No schooling 54.6 54.0 44.5 35.4 47.4 Primary 22.1 21.7 31.9 35.0 27.6 Secondary 18.3 19.5 18.3 24.7 20.0 Higher secondary and above 5.0 4.8 5.3 4.9 5.0 Dwelling ownership status Own 95.2 97.6 98.0 95.6 96.7 Rent 0.3 0.2 0.8 0.4 0.4 Allowed to reside 4.5 2.2 1.2 4.0 2.9 House type Pucca 1.5 1.0 2.8 0.6 1.5 Semi pucca 3.3 6.2 7.0 20.1 8.9 Kacha (but tin roof) 88.4 76.6 62.1 60.9 72.1 Kacha (thatch roof) 6.8 16.2 28.1 18.4 17.5 Latrine type Pucca 11.1 9.2 13.3 6.5 10.1 Slab 37.0 22.5 47.8 39.8 36.9 Kutcha 48.6 67.8 38.4 53.3 51.8 Other 3.3 0.5 0.5 0.4 1.2 Drinking water source Tap inside/outside home 0.5 0.3 1.1 0.2 0.6 Tube well 93.8 99.2 85.9 92.9 92.8 Dug well 0.4 0.3 5.3 -- 1.6 Pond/canal 5.3 0.2 7.7 6.9 5.0 Source: BIDS Survey (2004). 100 World Bank Working Paper Table A1.4. Annual Household Income by Source (average Tk per household) Source Dhaka Rajshahi Chittagong Khulna All Agriculture 21,096 28,652 31,498 25,546 26,797 Crop 9,225 13,090 13,469 10,590 11,644 Non-crop 9,359 9,846 14,212 10,720 11,092 Wages 2,512 5,716 3,817 4,236 4,061 Non-agriculture 45,510 19,226 48,210 24,544 34,832 Processing/trade/miscellaneous 15,824 7,277 13,109 8,828 11,349 Wages 6,191 2,442 3,420 3,141 3,810 Transport 3,042 3,047 3,681 2,786 3,156 Salaries/allowances/pensions 5,900 3,079 6,807 4,897 5,205 Rent (including land mortgages) 2,556 1,915 1,454 1,357 1,825 Remittances 11,997 1,466 19,739 3,535 9,487 Total 66,606 47,878 79,708 50,090 61,629 Source: BIDS Survey (2004). Table A1.5. Household Distribution by Income Source (percent) Source Dhaka Rajshahi Chittagong Khulna All Agriculture 94.5 96.8 95.5 95.4 95.6 Crop 64.8 72.3 68.6 65.9 68.0 Non-crop 91.2 89.2 93.3 90.7 91.1 Wages 23.5 43.7 18.9 31.2 29.1 Non-agriculture 88.2 75.3 78.9 83.6 81.4 Processing/trade/miscellaneous 35.0 42.8 34.2 41.4 38.2 Wages 29.9 19.2 14.8 21.7 21.3 Transport 10.0 14.3 7.3 13.1 11.1 Salaries/allowances/pensions 15.3 8.8 19.1 25.4 17.0 Rent (including land mortgages) 13.3 11.8 5.2 12.4 10.5 Remittances 27.7 11.0 28.6 13.3 20.2 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 101 Table A1.6. Household Consumption Expenditure by Budget Item (Tk per year) Item Dhaka Rajshahi Chittagong Khulna All Food 34,511 22,606 39,706 25,693 30,893 Clothing/footwear 3,679 2,374 3,286 2,656 3,012 Consumer durables 5,450 3,375 4,921 3,527 4,347 Miscellaneous 12,264 7,783 14,473 9,013 10,986 All 55,904 36,138 62,386 40,889 49,238 Source: BIDS Survey (2004). Table A1.7. Household Distribution by Type of Energy Consumption (percent) Energy Type Dhaka Rajshahi Chittagong Khulna All Biomass 99.8 99.5 98.6 100.0 99.5 Fuelwood 85.1 67.2 95.8 88.7 84.3 Tree leaves 81.4 72.0 61.6 91.8 76.1 Crop residue* 81.4 93.2 53.6 75.4 75.5 Dung cake/stick 56.9 72.3 29.5 64.8 55.2 Sawdust 0.7 0.8 0.3 1.6 0.8 Non-biomass 100.0 96.5 100.0 100.0 99.1 Candle 2.5 0.8 10.6 4.4 4.7 Kerosene 99.0 91.5 98.4 100.0 97.2 Natural gas -- -- 0.9 -- 0.3 LPG/LNG -- 0.2 1.1 -- 0.3 Grid electricity 43.8 20.3 38.9 10.8 29.0 Solar PV 0.3 -- -- 1.5 0.4 Storage cell 1.3 -- 0.8 0.9 0.8 Dry-cell battery 39.1 50.7 42.3 61.3 48.0 Source: BIDS Survey (2004). * Also includes crop waste and weeds. 102 World Bank Working Paper Table A1.8. Annual Household Energy Consumption: All Divisions (average per household) Heating Energy Type All Uses Cooking Parboiling Other Cooling Lighting Amusement Biomass (kg) Fuelwood 1,186.21 1,064.84 28.60 92.77 -- -- -- Tree leaves 501.51 470.67 29.99 0.85 -- -- -- Crop residue 708.18 538.86 164.41 2.72 -- -- -- Dung cake/stick 523.90 503.68 16.07 4.16 -- -- -- Sawdust 8.40 8.36 0.02 0.02 -- -- -- Non-biomass Candle (piece) 15.86 -- -- -- -- 15.86 -- Kerosene (liter) 28.98 1.76 -- 0.07 -- 27.16 -- Natural gas (Tk) 9.59 9.59 -- -- -- -- -- LPG/LNG (liter) 0.05 0.05 -- -- -- -- -- Grid electricity (kWh) 143.83 0.25 -- 4.00 49.50 80.74 9.34 Solar PV (kWh) 0.53 -- -- -- 0.04 0.48 0.01 Storage cell (kWh) 0.55 -- -- -- -- 0.14 0.41 Dry-cell battery (piece) 15.01 -- -- -- -- -- -- Source: BIDS Survey (2004). Note: Electricity use for motive power is negligible (0.28 kWh/ household/year). Only 2 households used briquettes (Chittagong), but only for 1­2 months. Only 1 household used charcoal, but in negligible quantities. Only 2 households used biogas. None used generator electricity. All categories include non-users so figures may vary due to the percentage of households not using a fuel. However, at the national level they should be representative for all households. Restoring Balance: Bangladesh's Rural Energy Realities 103 Table A1.8.1. Annual Household Energy Consumption: Dhaka (average per household) Heating Energy Type All Uses Cooking Parboiling Other Cooling Lighting Amusement Biomass (kg) Fuelwood 1,022.74 972.05 44.56 6.13 -- -- -- Tree leaves 536.39 495.65 38.02 2.72 -- -- -- Crop residue 619.10 501.42 116.21 1.47 -- -- -- Dung cake/stick 484.26 456.77 22.35 5.14 -- -- -- Sawdust 3.19 3.19 -- -- -- -- -- Non-biomass Candle (piece) 1.87 -- -- -- -- 1.87 -- Kerosene (liter) 26.02 3.38 -- -- -- 22.63 -- Natural gas (Tk) -- -- -- -- -- -- -- LPG/LNG (liter) -- -- -- -- -- -- -- Grid electricity (kWh) 232.50 0.70 -- 6.48 95.21 114.07 16.04 Solar PV (kWh) 0.66 -- -- -- 0.15 0.48 0.03 Storage cell (kWh) 1.52 -- -- -- -- 0.56 0.96 Dry-cell battery (piece) 9.92 -- -- -- -- -- -- Source: BIDS Survey (2004). Table A1.8.2. Annual Household Energy Consumption: Rajshahi (average per household) Heating Energy Type All Uses Cooking Parboiling Other Cooling Lighting Amusement Biomass (kg) Fuelwood 480.05 454.31 25.74 -- -- -- -- Tree leaves 365.47 348.10 17.07 0.31 -- -- -- Crop residue 847.10 723.86 118.47 4.78 -- -- -- Dung cake/stick 737.40 708.79 26.90 1.71 -- -- -- Sawdust 8.70 8.62 0.08 -- -- -- -- Non-biomass Candle (piece) 0.38 -- -- -- -- 0.38 -- Kerosene (liter) 22.08 0.68 -- 0.01 -- 21.39 -- Natural gas (Tk) -- -- -- -- -- -- -- LPG/LNG (liter) 0.02 0.02 -- -- -- -- -- Grid electricity (kWh) 64.10 0.30 -- 0.41 20.84 38.77 3.78 Solar PV (kWh) -- -- -- -- -- -- -- Storage cell (kWh) -- -- -- -- -- -- -- Dry-cell battery (piece) 15.50 -- -- -- -- -- -- Source: BIDS Survey (2004). 104 World Bank Working Paper Table A1.8.3. Annual Household Energy Consumption: Chittagong (average per household) Heating Energy Type All Uses Cooking Parboiling Other Cooling Lighting Amusement Biomass (kg) Fuelwood 2,069.84 1,714.93 18.70 336.21 -- -- -- Tree leaves 386.68 378.17 8.24 0.27 -- -- -- Crop residue 454.007 369.07 84.80 0.20 -- -- -- Dung cake/stick 235.95 232.82 2.42 0.70 -- -- -- Sawdust 4.50 4.50 -- -- -- -- -- Non-biomass Candle (piece) 51.94 -- -- -- -- 51.94 -- Kerosene (liter) 36.01 2.15 -- 0.06 -- 33.81 -- Natural gas (Tk) 35.81 35.81 -- -- -- -- -- LPG/LNG (liter) 0.15 0.15 -- -- -- -- -- Grid electricity (kWh) 226.15 -- -- 7.31 65.68 139.12 14.04 Solar PV (kWh) -- -- -- -- -- -- -- Storage cell (kWh) 0.25 -- -- -- -- -- 0.25 Dry-cell battery (piece) 15.57 -- -- -- -- -- -- Source: BIDS Survey (2004). Table A1.8.4. Annual Household Energy Consumption: Khulna (average per household) Heating Energy Type All Uses Cooking Parboiling Other Cooling Lighting Amusement Biomass (kg) Fuelwood 1,107.29 1,076.16 25.75 5.39 -- -- -- Tree leaves 746.18 685.41 60.70 0.07 -- -- -- Crop residue 941.29 575.80 360.71 4.78 -- -- -- Dung cake/stick 670.05 647.03 13.23 9.78 -- -- -- Sawdust 18.36 18.28 -- 0.08 -- -- -- Non-biomass Candle (piece) 6.07 -- -- -- -- 6.07 -- Kerosene (liter) 31.60 0.70 -- 0.2 -- 30.70 -- Natural gas (Tk) -- -- -- -- -- -- -- LPG/LNG (liter) -- -- -- -- -- -- -- Grid electricity (kWh) 37.40 -- -- 1.32 11.69 21.84 2.55 Solar PV (kWh) 1.57 -- -- -- -- 1.57 -- Storage cell (kWh) 0.42 -- -- -- -- -- 0.42 Dry-cell battery (piece) 19.40 -- -- -- -- -- -- Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 105 Table A1.9. Average Energy Consumption for Agriculture and Transportation (owner­users only) Agriculture Energy Type Irrigation pump Power tiller/tractor Thresher Transport Electricity (kWh) 3643.47 (9) -- 78.13 (17) -- Diesel (liter) 347.95 (103) 238.24 (35) -- 13,003.55 (14) Source: BIDS Survey (2004). Note: Figures in parentheses indicate the number of households that own and operate the relevant equipment. Table A1.10. Energy Consumption by Business Type Home Enterprise Village Enterprise Institution Users Energy Users Energy Users Energy (% total use/unit/ (% total use/unit/ (% total use/unit/ Energy Type units) year1 units) year1 units) year1 Biomass (kg) 26.3 12.3 10.3 Fuelwood 19.7 395.12 7.3 311.46 9.0 657.00 Tree leaves 7.3 74.54 0.6 2.11 -- -- Crop residue2 15.3 277.62 2.0 949.09 3.2 5.67 Dung cake/stick 5.1 21.64 3.5 149.00 1.9 8.38 Sawdust 2.2 77.52 1.2 6.18 -- -- Briquette -- -- 0.6 12.81 -- -- Charcoal 1.5 2.85 1.5 3.95 -- -- Non-biomass 72.3 98.5 74.4 Candle (piece) 3.6 5.26 21.1 18.16 45.5 100.87 Kerosene (liter) 55.5 13.52 78.9 84.04 37.8 18.66 Diesel/petrol (liter) 5.1 45.59 4.4 60.50 1.3 2.35 Grid electricity (kWh) 25.5 194.68 51.2 360.04 42.3 439.97 Storage cell (kWh) 0.7 3.15 2.6 12.00 7.1 39.00 Dry cell battery (piece) 2.9 0.83 23.7 6.24 12.8 3.50 All 79.6 98.5 75.0 Source: BIDS Survey (2004). 1 Average over all units. 2 Includes crop waste and weeds. 106 World Bank Working Paper Table A1.11. Household Distribution of Cooking Stove Ownership (percent) Stove Type Dhaka Rajshahi Chittagong Khulna All Divisions Clay (fixed) 99.0 99.7 98.6 100.0 99.3 Clay (portable) 30.7 31.7 3.0 11.9 19.2 Kerosene 6.5 1.0 0.8 1.1 2.3 Gas 0.2 -- 1.3 -- 0.4 Electric heater 0.3 0.2 0.2 -- 0.2 Source: BIDS Survey (2004). Table A1.12. Household Ownership of Cooking Stoves by Type (number per 100 households) Stove Type Dhaka Rajshahi Chittagong Khulna All Divisions Clay (fixed) 200.0 172.3 182.8 174.8 182.7 Clay (portable) 33.2 32.8 3.9 12.6 20.5 Kerosene 7.1 1.0 0.8 1.3 2.5 Gas 0.1 -- 1.3 -- 0.4 Electric heater 0.3 0.2 1.7 -- 0.2 All 240.7 206.3 190.5 188.7 206.3 Source: BIDS Survey (2004). Table A1.13. Household Distribution of Number of Meals and Cooking Times (percent distribution) Number of Meals Cooking Times Dhaka Rajshahi Chittagong Khulna All Divisions 4 2 0.5 -- 0.5 1.6 0.6 4 3 8.6 -- 0.3 -- 0.1 4 4 -- -- 0.2 -- -- 3 1 -- 3.3 8.6 32.5 12.8 3 2 77.4 81.0 47.6 63.1 67.1 3 3 11.6 12.7 38.9 2.2 17.0 2 2 0.7 1.6 0.3 0.4 0.8 2 2 1.2 1.2 3.6 0.2 1.6 1 1 -- -- -- -- -- All -- 0.2 -- -- -- Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 107 Table A1.14. Hours of Operating Cooking Stoves by Type (daily) Stove Type Dhaka Rajshahi Chittagong Khulna All Divisions Clay (fixed) 2.97 2.96 4.01 2.93 3.23 Clay (portable) 1.42 2.63 2.77 1.39 2.01 All 3.36 3.79 4.07 3.06 3.59 Source: BIDS Survey (2004). Note: Kerosene stoves and electric heaters are mainly for occasional use, but more information is needed; data on gas stoves are lacking. Table A1.15. Household Ownership of Lighting Appliances (number per 100 households) Appliance Type Dhaka Rajshahi Chittagong Khulna All Divisions Grid electrified households Kupi/cherag 153.0 162.3 212.4 161.0 176.7 Hurricane/lantern 80.3 77.0 95.6 81.4 85.3 Petromax 1.1 1.6 -- -- 0.7 Light bulbs 310.6 315.6 414.5 308.5 348.6 Tube lights 36.7 18.9 77.9 20.3 47.0 Charger (include charger torch) -- 0.8 19.3 1.7 7.2 Non-electrified and off-grid electrified households Kupi/cherag 162.5 160.5 210.2 196.7 182.8 Hurricane/lantern 89.7 90.2 98.2 84.7 90.3 Petromax 0.6 0.8 -- 0.4 0.5 Light bulbs -- -- -- -- -- Tube lights 2.9* -- -- 6.7* 2.5* Charger (include charger torch) -- -- 0.5 -- 0.1 All Households Kupi/cherag 158.4 160.8 211.1 192.9 181.0 Hurricane/lantern 85.6 87.5 97.2 84.3 88.9 Petromax 0.8 1.0 -- 0.4 0.5 Light bulbs 136.0 64.2 161.3 33.2 101.2 Tube lights 17.7 3.8 30.3 8.2 15.4 Charger (include charger torch) -- 0.2 7.8 0.2 2.2 Source: BIDS Survey (2004). * It should be noted that 11 households without an electricity connection have solar panels (10) biogas plant (1), which use tube lights. 108 World Bank Working Paper Table A1.16. Hours of Running Lighting Appliances (cases per day) Appliance Type Dhaka Rajshahi Chittagong Khulna All Divisions Kupi/cherag 1.58 (583) 2.05 (582) 2.24 (623) 2.18 (546) 2.01 (2334) Hurricane/lantern 3.32 (375) 4.15 (390) 2.97 (470) 3.63 (347) 3.49 (1582) Petromax 0.38 (4) 0.22 (4) -- 6.90 (2) 1.62 (10) Light bulbs 3.80 (259) 4.09 (121) 3.83 (247) 3.85 (54) 3.86 (681) Tube lights 3.54 (55) 4.65 (13) 3.91 (88) 3.67 (15) 3.82 (171) Torch/charger -- 0.50 (1) 1.74 (41) 2.00 (1) 1.71 (43) All 2.62 (1276) 3.03 (1111) 2.83 (1469) 2.83 (965) 2.82 (4821) Source: BIDS Survey (2004). Note: Figures in parentheses indicate number of cases. Table A1.17. Household Ownership of Plug-in Electric Appliances (number per 100 households) Appliance Type Dhaka Rajshahi Chittagong Khulna All Divisions Radio/cassette recorder 16.7 4.5 195.3 5.5 118.4 VCR/VCP/television 23.1 6.7 15.8 4.7 12.8 Electric cooling fan 50.1 17.3 56.1 8.0 33.8 Other 9.6 1.2 21.1 1.5 8.7 All 99.5 29.7 288.3 19.7 173.7 Source: BIDS Survey (2004). Table A1.18. Hours of Running the Plug-in Type of Electric Appliances (daily) Appliance Type Dhaka Rajshahi Chittagong Khulna All Divisions Radio/cassette recorder 2.06 (97) 2.95 (26) 2.41 (118) 2.84 (30) 2.38 (271) VCR/VCP/television 3.06 (133) 3.55 (41) 3.76 (101) 3.21 (26) 3.39 (300) Electric cooling fan 7.21 (163) 7.09 (57) 7.75 (177) 8.12 (25) 7.47 (422) Other 4.58 (53) 0.35 (5) 2.81 (129) 0.52 (8) 3.36 (182) All 4.54 (446) 4.87 (129) 4.57 (525) 4.22 (89) 4.62 (1175) Source: BIDS Survey (2004). Note: Figures in parentheses indicate number of cases. Restoring Balance: Bangladesh's Rural Energy Realities 109 Table A1.19. Ownership of Irrigation Pumps and Power Tillers/Tractors (percent households) Pump/Vehicle Type Dhaka Rajshahi Chittagong Khulna All Divisions Irrigation pump (electric) -- 1.3 0.2 -- 0.4 Irrigation pump (diesel) 1.8 12.2 0.5 4.0 4.6 Power tiller/tractor (diesel) 1.3 3.8 -- 1.6 1.7 Source: BIDS Survey (2004). Table A1.20. Hours of Running Pumps and Power Tillers/Tractors (annual hours per unit) Pump/Vehicle Type Dhaka Rajshahi Chittagong Khulna All Divisions Irrigation pump (electric) -- 1,382 1,050 -- 1,345 Irrigation pump (diesel) 441 404 341 283 381 Power tiller/tractor (diesel) 548 239 -- 856 440 Source: BIDS Survey (2004). Table A1.21. Cost of Household Energy Consumption: All Divisions (Tk per household per year) Heating Energy Source All Uses Cooking Parboiling Other Cooling Lighting Amusement Biomass 3,798.50 3,379.57 242.17 176.76 -- -- -- Fuelwood 1,962.25 1,749.59 44.78 167.88 -- -- -- Tree leaves 470.40 440.22 29.14 1.04 -- -- -- Crop residue* 641.23 491.80 147.08 2.35 -- -- -- Dung cake/stick 716.34 689.72 21.15 5.47 -- -- -- Sawdust 8.28 8.24 0.02 0.02 -- -- Non-biomass 1,303.26 65.24 -- 10.56 134.33 -- -- Candle 7.83 -- -- -- -- 7.83 -- Kerosene 607.67 37.04 -- 1.38 -- 569.25 -- Natural gas 9.59 9.59 -- -- -- -- -- LPG/LNG 18.02 18.02 -- -- -- -- -- Grid electricity 487.69 0.59 -- 9.18 134.26 311.89 31.77 Solar PV 1.08 -- -- -- 0.07 0.99 0.02 Storage cell 10.73 -- -- -- -- 2.75 7.98 Dry-cell battery 160.65 -- -- -- -- -- -- All 5,101.76 3,444.81 242.17 187.32 134.33 -- -- Source: BIDS Survey (2004). * Also includes crop waste and weeds. 110 World Bank Working Paper Table A1.21.1. Cost of Household Energy Consumption: Dhaka (Tk per household per year) Heating Energy Source All Uses Cooking Parboiling Other Cooling Lighting Amusement Biomass 3,802.63 3,475.16 305.32 22.15 -- -- -- Fuelwood 1,610.82 1,530.99 70.18 9.65 -- -- -- Tree leaves 712.86 658.72 50.53 3.61 -- -- -- Crop residue 822.78 666.39 154.44 1.95 -- -- -- Dung cake/stick 653.75 616.64 30.17 6.94 -- -- -- Sawdust 2.42 2.42 -- -- -- -- -- Non-biomass 1,414.24 73.67 -- 13.68 226.97 -- -- Candle 7.64 -- -- -- -- 7.64 -- Kerosene 560.10 72.80 -- 0.17 -- 487.13 -- Natural gas -- -- -- -- -- -- -- LPG/LNG -- -- -- -- -- -- -- Grid electricity 710.35 0.87 -- 13.51 226.67 414.83 54.47 Solar PV 1.35 -- -- -- 0.30 0.99 0.06 Storage cell 18.33 -- -- -- -- 4.69 13.64 Dry-cell battery 116.47 -- -- -- -- -- -- All 5,216.87 3,548.83 305.32 35.83 226.97 -- -- Source: BIDS Survey (2004). Table A1.21.2. Cost of Household Energy Consumption Rajshahi (Tk per household per year) Heating Energy Source All Uses Cooking Parboiling Other Cooling Lighting Amusement Biomass 2,541.51 2,361.38 174.11 6.02 -- -- Fuelwood 685.03 648.30 36.73 -- -- -- -- Tree leaves 279.58 266.29 13.06 0.23 -- -- -- Crop residue 648.03 553.75 90.63 3.65 -- -- -- Dung cake/stick 921.74 885.98 33.62 2.14 -- -- -- Sawdust 7.13 7.06 0.07 -- -- -- -- Non-biomass 906.26 22.03 -- 1.66 74.91 -- -- Candle 0.88 -- -- -- -- 0.88 -- Kerosene 474.24 14.56 -- 0.13 -- 459.55 -- Natural gas -- -- -- -- -- -- -- LPG/LNG 6.00 6.00 -- -- -- -- -- Grid electricity 284.24 1.47 -- 1.53 74.91 193.62 12.71 Solar PV -- -- -- -- -- -- -- Storage cell 4.92 -- -- -- -- 1.26 3.66 Dry-cell battery 135.98 -- -- -- -- -- -- All 3,447.77 2,383.20 174.11 7.68 74.91 -- -- Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 111 Table A1.21.3. Cost of Household Energy Consumption Chittagong (Tk per household per year) Heating Energy Source All Uses Cooking Parboiling Other Cooling Lighting Amusement Biomass 4,901.92 4,177.90 111.01 613.01 -- -- -- Fuelwood 3,762.96 3,117.74 33.99 611.23 -- -- -- Tree leaves 299.67 293.08 6.38 0.21 -- -- -- Crop residue 351.90 286.03 65.72 0.15 -- -- -- Dung cake/stick 478.97 472.63 4.92 1.42 -- -- -- Sawdust 8.42 8.42 -- -- -- -- -- Non-biomass 1,793.87 141.15 -- 19.12 187.15 -- -- Candle 17.46 -- -- -- -- 17.46 -- Kerosene 732.15 43.63 -- 1.25 -- 687.27 -- Natural gas 35.81 35.81 -- -- -- -- -- LPG/LNG 61.71 61.71 -- -- -- -- -- Grid electricity 761.21 -- -- 17.87 187.15 507.43 48.76 Solar PV -- -- -- -- -- -- -- Storage cell 7.07 -- -- -- -- 1.81 5.26 Dry-cell battery 178.46 -- -- -- -- -- -- All 6,695.79 4,319.05 111.01 632.13 187.15 -- -- Source: BIDS Survey (2004). Table A1.21.4. Cost of Household Energy Consumption: Khulna (Tk per household per year) Heating Energy Source All Uses Cooking Parboiling Other Cooling Lighting Amusement Biomass 3,881.41 3,456.80 400.33 24.28 -- -- -- Fuelwood 1,644.33 1,598.09 38.24 8.00 -- -- -- Tree leaves 611.87 562.04 49.77 0.06 -- -- -- Crop residue 771.86 472.16 295.78 3.92 -- -- -- Dung cake/stick 837.56 808.79 16.54 12.23 -- -- -- Sawdust 15.79 15.72 -- 0.07 -- -- -- Non-biomass 1,042.79 14.61 -- 6.88 35.77 -- -- Candle 4.42 -- -- -- -- -- -- Kerosene 660.73 14.61 -- 4.22 -- 641.90 -- Natural gas -- -- -- -- -- -- -- LPG/LNG -- -- -- -- -- -- -- Grid electricity 145.98 -- -- 2.66 35.77 99.73 7.82 Solar PV 3.22 -- -- -- -- 3.22 -- Storage cell 12.98 -- -- -- -- 3.32 9.66 Dry-cell battery 215.46 -- -- -- -- -- -- All 4,924.20 3,471.41 400.33 31.16 35.77 -- -- Source: BIDS Survey (2004). 112 World Bank Working Paper Table A1.22. Household Biomass Energy Behavior: All Divisions (percent distribution) Energy Source Production Gathering Purchasing Any Combination None Fuelwood 9.6 24.6 17.8 10.3 37.7 Tree leaves 30.3 16.7 0.4 29.5 23.1 Crop residue* 36.0 9.8 4.7 27.0 22.5 Dung cake/stick 0.3 0.2 6.7 2.1 90.7 Sawdust -- -- 0.9 -- 99.1 Source: BIDS Survey (2004). *Also includes crop wastes and weeds; but mainly it is crop residue. Table A1.22.1. Household Biomass Energy Behavior: Dhaka (percent distribution) Energy Source Production Gathering Purchasing Any Combination None Fuelwood 10.1 18.1 24.9 11.3 35.6 Tree leaves 34.8 16.7 0.2 30.2 18.1 Crop residue 36.1 5.3 10.8 31.2 16.6 Dung cake/stick 0.2 -- 10.4 3.5 85.9 Sawdust -- -- 0.8 -- 99.2 Source: BIDS Survey (2004). Table A1.22.2. Household Biomass Energy Behavior: Rajshahi (percent distribution) Energy Source Production Gathering Purchasing Any Combination None Fuelwood 6.0 30.7 6.8 2.3 54.2 Tree leaves 17.8 24.2 -- 30.3 27.7 Crop residue 38.0 14.2 3.2 42.6 2.0 Dung cake/stick 0.5 0.2 2.8 1.4 95.1 Sawdust -- -- 0.8 -- 99.2 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 113 Table A1.22.3. Household Biomass Energy Behavior: Chittagong (percent distribution) Energy Source Production Gathering Purchasing Any Combination None Fuelwood 12.5 33.4 20.3 13.4 20.3 Tree leaves 27.3 10.5 0.6 23.8 37.8 Crop residue 32.7 10.8 1.1 9.0 46.4 Dung cake/stick 0.5 0.2 2.8 1.4 95.2 Sawdust -- -- 0.3 -- 99.7 Source: BIDS Survey (2004). Table A1.22.4. Household Biomass Energy Behavior: Khulna (percent distribution) Energy Source Production Gathering Purchasing Any Combination None Fuelwood 9.7 15.0 19.2 14.2 42.0 Tree leaves 42.3 15.9 0.9 34.5 6.4 Crop residue 37.8 8.9 3.8 26.1 23.4 Dung cake/stick 0.4 0.2 2.4 1.4 95.6 Sawdust -- -- 1.8 -- 98.2 Source: BIDS Survey (2004). Table A1.23. Household Consumption of Selected Biomass Energies: All Divisions (per year) Household % Distribution by Source Energy Type Consumption (kg/yr) Own Production Gathered Purchased Fuelwood 1,186.21 10.96 49.76 39.28 Tree leaves 501.51 52.23 46.33 1.44 Crop residue 708.18 68.55 24.67 6.78 Dung cake/stick 523.90 72.22 5.25 22.53 Sawdust 8.40 -- -- 100.00 All 2,928.20 42.89 35.00 22.11 Source: BIDS Survey (2004). 114 World Bank Working Paper Table A1.23.1. Household Consumption of Selected Biomass Energies: Dhaka (per year) Household % Distribution by Source Energy Type Consumption (kg/yr) Own Production Gathered Purchased Fuelwood 1,022.74 6.30 26.86 66.84 Tree leaves 536.39 55.54 43.89 0.57 Crop residue 619.10 72.16 15.02 12.82 Dung cake/stick 484.26 23.19 9.23 67.58 Sawdust 3.19 -- -- 100.00 All 2,665.68 34.56 24.30 41.14 Source: BIDS Survey (2004). Table A1.23.2. Household Consumption of Selected Biomass Energies: Rajshahi (per year) Household % Distribution by Source Energy Type Consumption (kg/yr) Own Production Gathered Purchased Fuelwood 480.05 7.32 55.42 37.26 Tree leaves 365.47 40.96 58.90 0.14 Crop residue 847.10 67.47 28.28 4.25 Dung cake/stick 737.40 24.98 17.12 57.90 Sawdust 8.70 -- -- 100.00 All 2,438.72 38.57 34.73 26.70 Source: BIDS Survey (2004). Table A1.23.3. Household Consumption of Selected Biomass Energies: Chittagong (per year) Household % Distribution by Source Energy Type Consumption (kg/yr) Own Production Gathered Purchased Fuelwood 2,069.84 14.02 61.90 24.08 Tree leaves 386.68 57.77 41.17 1.06 Crop residue 454.07 69.36 28.85 1.79 Dung cake/stick 235.95 82.77 3.58 13.65 Sawdust 4.50 -- -- 100.00 All 3,151.04 32.49 50.14 17.37 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 115 Table A1.23.4. Household Consumption of Selected Biomass Energies: Khulna (per year) Household % Distribution by Source Energy Type Consumption (kg/yr) Own Production Gathered Purchased Fuelwood 1,107.29 7.61 32.44 59.95 Tree leaves 746.18 55.51 40.33 4.16 Crop residue 941.29 66.58 25.66 7.76 Dung cake/stick 670.05 92.49 1.30 6.21 Sawdust 18.36 -- -- 100.00 All 3,483.17 50.10 26.14 23.77 Source: BIDS Survey (2004). Table A1.24. Operational Land and Household Ownership of Gardens Characteristic Dhaka Rajshahi Chittagong Khulna All Divisions Operational land Households with 53.23 69.50 61.72 59.49 61.02 cultivation (percent) Operational area/household 214.58 376.50 321.76 276.69 276.43 (decimal)* Garden Household ownership 9.29 13.50 20.63 43.61 21.25 (percent) Area under garden/household 78.13 23.38 202.06 26.49 77.31 (decimal)* Source: BIDS Survey (2004). * Refers to ownership households; 1 decimal = 50 square yards. Table A1.25. Household Ownership of Trees and Bamboo Characteristic Dhaka Rajshahi Chittagong Khulna All Divisions Trees % household ownership 89.05 83.67 88.75 89.60 87.75 Number of mature trees/ 37.18 22.24 79.65 70.19 52.83 household* Number of immature trees/ 39.15 34.36 256.27 111.56 113.73 household* Bamboo groves % household own 47.76 44.33 38.28 44.89 43.71 Number of mature bamboo/ 119.05 90.77 474.22 69.83 183.53 household* Number of immature bamboo/ 74.26 93.92 499.03 79.95 180.19 household* Source: BIDS Survey (2004). * Refers to ownership households. 116 World Bank Working Paper Table A1.26. Household Ownership of Cattle Characteristic Dhaka Rajshahi Chittagong Khulna All Divisions Cows % household ownership 39.30 40.67 53.44 46.35 45.04 Number of cows/household* 2.63 2.83 2.77 2.48 2.68 Buffaloes % household ownership 0.17 0.83 0.31 1.46 0.67 Number of buffaloes/household* 1.00 2.00 2.00 3.50 2.69 Source: BIDS Survey (2004). * Refers to ownership households. Table A1.27. Household Distribution by Duration of Electricity Connection (percent distribution) Duration of Connection Dhaka Rajshahi Chittagong Khulna All Divisions No connection 56.2 79.7 61.1 89.2 71.0 With connection (months) 43.8 20.3 38.9 10.8 29.0 01--06 4.6 1.7 2.5 0.6 2.4 07--12 7.3 1.3 3.4 0.7 3.3 13--24 9.6 2.0 5.6 2.6 5.0 25--36 4.5 2.3 5.6 2.9 3.9 37--60 6.3 4.5 4.8 1.6 4.4 61 or more 10.0 7.8 16.4 2.0 9.3 Date not reported 1.5 0.7 0.5 0.4 0.7 Source: BIDS Survey (2004). Table A1.28. Electricity Sources (percent of connected households) Source Dhaka Rajshahi Chittagong Khulna All Divisions Palli Bidyut Samity (PBS) 70.8 84.4 71.5 72.9 73.6 Neighbor with PBS connection 28.4 15.6 17.7 22.0 21.8 Power Development Board (PDB) 0.4 -- 5.6 -- 2.2 Neighbors with PDB connection -- -- 5.2 1.7 2.0 Other (own or local generation) 0.4 -- -- 3.4 0.4 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 117 Table A1.29. Basis of Charging for Electricity (percent of connected households) Basis Dhaka Rajshahi Chittagong Khulna All Divisions Kilowatt hours consumed 67.8 81.1 75.1 62.7 72.3 Number of electric appliances/ 15.9 6.6 5.2 18.6 10.7 bulbs/tubes Monthly fixed charge 7.6 7.4 18.9 10.2 11.8 No charge 1.5 0.8 0.4 -- 0.9 Other 2.7 3.3 0.4 6.8 2.3 Not reported 4.5 0.8 -- 1.7 2.0 Source: BIDS Survey (2004). Table A1.30. Household Distribution by Electricity Supply Problem (percent of connected households) Problem Dhaka Rajshahi Chittagong Khulna All Divisions Unscheduled power cuts 74.2 69.7 93.6 59.3 79.1 Daily 6.4 0.8 2.0 6.8 3.9 Weekly 10.2 27.1 4.0 20.3 11.8 Rarely 6.4 0.8 -- 10.2 3.5 Never 2.7 1.6 0.4 3.4 1.7 Not reported Voltage fluctuation 44.3 53.3 81.1 55.9 60.1 Daily 14.8 1.6 -- 3.4 6.2 Weekly 34.1 37.7 15.7 32.2 28.0 Rarely 4.2 5.7 2.8 5.1 4.0 Never 2.6 1.6 0.4 3.4 1.7 Not reported Source: BIDS Survey (2004). Table A1.31. Cost of Kerosene--Using Devices Monthly Repair Purchase Value/ Cost/Household Device Household (Tk) (Tk) Kupi/cherag 14.05 (2,332) 0.09 (2,327) Hurricane 86.09 (1,645) 2.66 (1,641) lantern Petromax 810.42 (12) 25.42 (12) Source: BIDS Survey (2004). Note: Figures in parentheses indicate number of cases. 118 World Bank Working Paper Table A1.32. Cost of Electric Bulbs and Tube Lights Item Annual Cost/Household Bulbs/tube lights bought (no.) 5.23 (648) Cost (Tk) 111.79 (648) Source: BIDS Survey (2004). Note: Figures in parentheses indicate number of cases. Table A1.33. Women's Time Use for Various Activities (daily hours per person) Activity Type Dhaka Rajshahi Chittagong Khulna All Divisions Crop processing 0.24 0.48 0.44 0.55 0.43 Collecting energy 0.22 0.48 0.41 0.31 0.36 Cooking/washing/ 6.26 5.81 6.34 6.18 6.15 feeding/eating Fetching water/washing 1.38 1.62 1.47 1.26 1.44 clothes Cleaning house/bathing 1.87 1.80 1.55 1.48 1.68 Childcare 1.27 1.03 1.29 1.17 1.19 Gainful work 0.83 0.70 0.97 0.54 0.77 Religious work 0.98 0.91 1.22 0.76 0.98 Study/reading 0.07 0.06 0.07 0.04 0.06 Watching TV/listening 1.41 1.37 1.24 1.43 1.36 to radio Visiting neighbors 0.95 1.07 0.77 0.95 0.93 Sleeping 8.19 8.46 8.01 8.61 8.31 Other 0.32 0.20 0.21 0.72 0.35 Source: BIDS Survey (2004). Table A1.34. Time Spent Making Clay Stoves (hours per stove) Stove Type Adult Male Adult Female Fixed clay 3.83 5.10 Portable burnt clay 6.05 6.88 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 119 Table A1.35. Time Spent on Clay Stove Maintenance Times per Month Repair Hours Total Monthly Stove Type Each Stove Is Repaired per Incident Repair Hours Fixed clay 4.80 0.54 2.59 Portable burnt clay 2.60 0.41 1.07 Source: BIDS Survey (2004). Table A1.36. Respondents' Attitude toward Electricity (percent distribution) Advantage/Disadvantage Rank­1 Rank­2 Rank­3 Advantage Electric light is superior to kerosene light 70.0 10.0 6.2 Easier to study by electric light than kerosene light 3.6 13.4 4.2 Children devote more time to study at night if electric light 6.5 16.0 12.2 People feel secure at night with electric light 5.3 26.2 18.3 Electric light is not as harmful to health as kerosene light 0.3 3.9 4.2 Running TV is easier by electricity than battery 2.4 5.7 8.5 Electricity costs less than battery to run TV 0.6 2.6 4.6 Hard to work at night without electricity 1.7 6.1 12.4 Easier to entertain guests at night if electricity -- 1.0 2.3 Electricity is important for water supply in our locality 1.1 1.0 2.0 Cooking with electricity causes no smoke 0.2 0.7 1.7 Life is easy with electricity 2.7 5.2 13.1 Disadvantage Children will waste reading time watching TV 16.3 12.1 5.8 Electricity often causes accident that ends in death 54.9 19.3 8.5 Cooking is not advantageous with electricity 4.9 8.7 7.5 Electricity is very expensive 9.3 24.8 19.4 Bulb/tube etc. is expensive 3.0 17.0 22.4 Electric supply is very irregular and voltage is often low 10.1 13.6 22.4 Source: BIDS Survey (2004). 120 World Bank Working Paper Table A1.37. Respondents' Attitude toward Fuelwood (percent distribution) Advantage/Disadvantage Rank­1 Rank­2 Rank­3 Advantage Fuelwood is easy to collect 40.9 22.0 13.3 Fuelwood is cheaper in the market 2.8 10.1 11.5 Fuelwood is obtained from own trees without any cost 25.3 26.1 12.4 Disadvantage Cooking with fuelwood is not advantageous 6.8 4.2 2.5 Fuelwood generates smoke that creates breathing problems 56.1 16.9 8.8 Cooking with fuelwood is harmful to health 3.7 20.5 13.6 Fuelwood is very expensive 12.8 19.2 8.9 Getting fuelwood in the market is difficult 1.5 6.4 7.7 Excessive use of fuelwood is the root cause of deforestation 13.0 19.2 29.0 Source: BIDS Survey (2004). Table A1.38. Respondents' Attitude toward Kerosene (percent distribution) Advantage/Disadvantage Rank­1 Rank­2 Advantage Cooking is easier with kerosene 22.0 12.4 Kerosene is not expensive for lighting 35.9 31.1 Kerosene is the best way for lighting for our family 28.6 21.2 Disadvantage Use of kerosene is very expensive for cooking 57.4 18.9 Foods cooked with kerosene are harmful to health 24.0 48.8 Kerosene is not easily available in the market 1.8 4.5 Source: BIDS Survey (2004). Table A1.39. Respondents' Attitude toward LPG/LNG (percent distribution) Advantage/Disadvantage Rank­1 Rank­2 Advantage Cooking is easier with LPG/LNG 36.4 1.2 Foods cooked with LPG/LNG are not harmful to health 1.1 19.4 Disadvantage Use of LPG/LNG is very expensive for cooking 25.3 6.0 LPG/LNG is not easily available in the market 8.0 15.5 LPG/LNG cooking devices are expensive 2.0 11.1 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 121 Table A1.40. Respondents' Knowledge of Price of Kerosene (percent distribution) Factor Dhaka Rajshahi Chittagong Khulna All Divisions Kerosene is cheaper 1.8 0.5 0.5 1.3 1.0 Price is justified 3.8 11.2 9.7 4.6 7.4 Kerosene is costly 93.9 88.2 89.7 93.8 91.3 Do not know 0.5 0.2 0.2 0.4 0.3 Source: BIDS Survey (2004). Table A1.41. Respondents' Reasons for Costliness of Kerosene (percent of those who reported kerosene as costly) Reason Dhaka Rajshahi Chittagong Khulna All Divisions Kerosene is taxed 3.4 7.0 1.4 1.2 3.2 No subsidy on kerosene 0.4 4.7 0.5 0.2 1.4 Iraq--U.S. war 38.0 12.7 12.9 33.5 24.2 Do not know 58.3 75.6 85.2 65.2 71.2 Source: BIDS Survey (2004). Table A1.42. Respondents' Knowledge of Price of Electricity (percent distribution) Factor Dhaka Rajshahi Chittagong Khulna All Divisions Electricity is cheaper 6.3 3.0 1.6 1.8 3.2 Price is justified 8.3 3.7 18.0 6.4 9.3 Kerosene is costly 70.5 54.8 68.4 51.5 61.6 Do not know 14.9 38.5 12.0 40.3 25.9 Source: BIDS Survey (2004). Table A1.43. Respondents' Reasons for Costliness of Electricity (percent of those who reported electricity as costly) Reason Dhaka Rajshahi Chittagong Khulna All Divisions Electricity is taxed 4.5 13.4 2.5 2.1 5.4 No subsidy on electricity 0.5 8.8 0.9 0.7 2.5 Other 74.1 70.5 87.9 82.6 79.0 Do not know 20.9 7.3 8.7 14.6 13.0 Source: BIDS Survey (2004). 122 World Bank Working Paper Table A1.44. Respondents' Knowledge of Renewable Energy (percent distribution) Whether Available WhetherInterested in the Market in Purchasing (% of those who (% of those who have heard about) have heard about) Renewable Energy Type Heard About Yes No Yes No Personal biogas plant 16.3 75.1 24.9 78.9 21.1 Improved stove 8.8 59.0 41.0 71.1 28.9 Solar cooker 3.2 47.3 52.7 59.7 40.3 Solar PV light 12.0 31.8 68.2 72.1 27.9 Solar PV lantern 0.9 55.0 45.0 55.0 45.0 Biogas light 3.4 82.7 17.3 67.1 32.9 Solar PV water heater 0.6 61.5 38.5 35.7 64.3 Solar PV water pump 0.6 60.0 40.0 33.3 66.7 Windmill water pump 0.7 68.7 31.3 37.5 62.5 Pressure cooker 25.0 9.0 91.0 57.5 42.5 Source: BIDS Survey (2004). Table A1.45. Respondents' Reasons for Interest in Purchasing Renewable Energy Devices (percent of those interested in purchasing) Reason Easily Can buy Renewable Energy Type Cheaper Available Prestige Advantage on Credit Other Personal biogas plant 2.5 -- 3.7 82.7 1.2 9.9 Improved stove -- -- -- 93.1 -- 6.9 Solar cooker -- -- 3.4 82.8 6.9 6.9 Solar PV light -- 2.5 5.1 54.4 30.4 7.6 Solar PV lantern -- -- 11.1 77.8 -- 11.1 Biogas light 3.7 -- 3.7 77.8 -- 14.8 Solar PV water heater -- -- 11.1 77.8 -- 11.1 Solar PV water pump -- -- -- 80.0 -- 20.0 Windmill water pump -- -- -- 80.0 10.0 10.0 Pressure cooker 0.4 2.8 4.3 82.3 -- 10.2 Source: BIDS Survey (2004). Restoring Balance: Bangladesh's Rural Energy Realities 123 Table A1.46. Respondents' Reasons for Disinterest in Purchasing Renewable Energy Devices (percent of those not interested in purchasing) Reason Not Known Unsure Cannot Where Lack of About Not Easily Buy on Renewable Energy Type Available Costly Money Utility Available Credit Other Personal biogas plant 10.3 9.6 23.0 25.9 4.6 2.1 24.5 Improved stove 22.0 12.9 12.9 42.4 3.0 0.8 6.0 Solar cooker 16.7 21.4 16.7 26.2 4.7 2.4 11.9 Solar PV light 7.4 22.3 29.2 13.4 1.0 2.0 24.7 Solar PV lantern 10.0 20.0 10.0 30.0 10.0 -- 20.0 Biogas light 4.0 4.0 34.0 22.0 2.0 2.0 32.0 Solar PV water heater -- -- 25.0 25.0 25.0 -- 25.0 Solar PV water pump -- -- 25.0 25.0 25.0 -- 25.0 Windmill water pump -- -- 25.0 25.0 25.0 -- 25.0 Pressure cooker 1.8 25.2 14.7 12.7 1.2 2.7 21.7 Source: BIDS Survey (2004). ANNEX 2 Statistical Models Chapter 3 Table A2.1. Tobit Estimate of Household Energy Demand (N = 2,388) Household Energy Demand Fuelwood Kerosene Diesel Electricity Explanatory Variable (kg/month) (liter/month) (liter/month) (kWh/month) Gender of household head 1.614 0.204 0.064 4.275 (1 = male, 0 = female) Age of household head (yrs.) ­0.051 0.005** 0.001 1.049 Maximum education adult 1.460** 0.019* 0.001 1.485 males in household (yrs.) Maximum education adult 0.759 0.017 ­0.001 2.382 females in household (yrs.) Log of household landholding 2.389** 0.105** 0.025** 4.640 (decimals)1 Log of household non-land 2.591** 0.031** 0.038** 4.448** asset (Tk) Energy price Fuelwood (Tk/kg) ­34.732** 0.321** ­0.003 ­0.119 Kerosene (Tk/liter) 5.951** ­0.009 0.002 0.534 Diesel (Tk/liter) ­8.807** ­0.106** 0.001 ­0.411 (continued ) 125 126 World Bank Working Paper Table A2.1. Tobit Estimate of Household Energy Demand (N = 2,388) (Continued) Household Energy Demand Fuelwood Kerosene Diesel Electricity Explanatory Variable (kg/month) (liter/month) (liter/month) (kWh/month) If village has electricity 14.211** ­0.597** ­0.015 32.782** Proportion irrigated area ­2.039 ­0.242** 0.031* 1.739** in village Whether village has any Paved road 7.282** ­0.033 ­0.025* 9.068 Primary school 9.818** ­0.157* 0.016 ­7.571 Health center ­19.016** 0.053 ­0.017 6.986 Market ­8.738** ­0.002 ­0.018 ­0.972 Commercial bank ­47.877** 0.442* 0.048 31.080 NGO 18.844** ­0.306* ­0.024 5.418 Village price Rice (Tk/kg) ­2.226 ­0.168** 0.002 ­3.040 Flour (Tk/kg) ­5.524** ­0.027 ­0.006 ­2.390 Potato (Tk/kg) ­4.242** 0.016 0.005 9.337* Lentil (Tk/kg) ­4.163** 0.006 0.005* ­2.321 Fish (Tk/kg) ­0.093 ­0.001** ­0.001* 0.173 Beef (Tk/kg) ­1.196** 0.001** 0.001 1.163 Chicken (Tk/kg) ­0.694** ­0.001* 0.0001 0.720 Soybean oil (Tk/liter) 0.210 ­0.019 ­0.004 ­3.139* Milk (Tk/liter) 2.618** 0.001 ­0.002 0.599 Sugar (Tk/kg) 3.173** 0.057** 0.001 ­2.329 Salt (Tk/kg) 10.741** 0.039 0.011 7.966 Village wage (Tk/day) Male 0.704** 0.008** ­0.001 0.192 Female ­0.913** ­0.008** 0.002 ­0.648 Child 0.488** ­0.005 0.001 0.294 Region2 Dhaka 21.673** ­0.079 ­0.066** ­87.485** Chittagong 84.843** 0.172 ­0.157** ­104.459** Khulna 27.633** 0.387** ­0.041* ­88.941** Constant 326.571** 4.702** ­0.870** 4.842 Log likelihood ­11,984.917 1,439.135 ­876.239 1,342.894 Source: BIDS Survey (2004). Note: ** and * represent significance levels of 5 percent (or higher) and 10 percent, respectively. 1 1 decimal = 1/100 acre. 2 Among the regions Rajshahi is the excluded category. Restoring Balance: Bangladesh's Rural Energy Realities 127 Chapter 4 It is difficult to identify the causal effect of access to alternate energy sources on household income and consumption. Since energy demand and household consumption/income are jointly determined, it is difficult to correlate them without imposing certain restrictions on the demand function. The restriction applied here is that energy demand is influenced by energy prices, which can affect household income or consumption indirectly through demand. It is assumed that energy prices are determined by supply and demand, without individual households having any significant effect. Thus, a two-stage, least-squared instrument variable was used for this estimation (see Tables A2.2 and A2.3), where demand for energy sources was determined in the first phase. Table A2.2. Effects of Energy Use on Household Outcomes by Use Type (N = 2,388) Household per Capita Consumption Household Income (Tk/month)1 Explanatory Variable (Tk/month) Farm Non-farm Total Gender of household head ­0.0440 0.598** ­1.049** 0.074 (1 = male, 0 = female) Age of household head (years) 0.0003 ­0.002 0.009** 0.001 Maximum education of 0.0140** ­0.023* 0.120** 0.024** adult males (years) Maximum education of 0.0005 ­0.022* 0.072** 0.021** adult females (years) Log of household landholding 0.0450** 0.309** ­0.160 0.066** (decimal)1 Log of non-land household 0.0110** 0.162** 0.060** 0.049** asset (Tk) Non-lighting energy use 0.1270** 0.410** 0.025 0.208** (kgoe/year) Lighting energy use 0.1360** ­0.010 0.387** 0.212** (klmn­hr/year) If village has electricity ­0.0130 ­0.141 0.903** ­0.058 Proportion irrigated area 0.0120 ­0.153 ­0.002 0.021 in village Whether village has any Paved road 0.0080 0.005 0.136 0.070** Primary school 0.0290 0.158* ­0.047 0.063* Health center ­0.0060 ­0.368** 0.017 ­0.063 Market ­0.0120 0.079 0.132 ­0.021 Commercial bank ­0.0890 0.095 ­0.506 ­0.158* NGO ­0.0780 ­0.020 0.235 0.081 (continued ) 128 World Bank Working Paper Table A2.2. Effects of Energy Use on Household Outcomes by Use Type (N = 2,388) (Continued) Household per Capita Consumption Household Income (Tk/month)1 Explanatory Variable (Tk/month) Farm Non-farm Total Village price Rice (Tk/kg) 0.0003 ­0.161** 0.210** 0.008 Flour (Tk/kg) ­0.0040 ­0.089** ­0.082 0.011 Potato (Tk/kg) 0.0110 ­0.013 ­0.026 ­0.005 Lentil (Tk/kg) ­0.0050 0.047** ­0.092** ­0.010* Fish (Tk/kg) 0.0010* ­0.004** ­0.003 0.001 Beef (Tk/kg) 0.0020* ­0.004 0.034** 0.007** Chicken (Tk/kg) ­0.0040** ­0.002 ­0.003 ­0.004** Soybean oil (Tk/liter) 0.0010 0.007 0.073** ­0.009* Milk (Tk/liter) 0.0090** 0.020 ­0.001 ­0.009* Sugar (Tk/kg) 0.0090* 0.028 ­0.063* ­0.001 Salt (Tk/kg) 0.0050 0.017 ­0.123* 0.010 Village wage (Tk/day) Male 0.0030** ­0.007** 0.013** 0.004** Female ­0.0020* ­0.001 ­0.009 ­0.003** Child 0.0001 ­0.002 ­0.003 0.0005 Region2 Dhaka ­0.0510 ­0.263* 1.207** 0.088 Chittagong ­0.1950** 0.255 0.192 0.145* Khulna ­0.1920** ­0.403** 0.961** ­0.028 Constant 5.0630** 1.158 ­5.955** 6.170** R-squared 0.2220 0.340 0.199 0.423 Source: BIDS Survey (2004). Note: ** and * represent significance levels of 5 percent (or higher) and 10 percent, respectively. Consumption and income variables are in log form. 1 1 decimal = 1/100 acre. 2 Among the regions, Rajshahi is the excluded category. Table A2.3. Effects of Energy Use on Household Outcomes by Energy Source (N = 2,388) Consumption Household Income (Tk/year) Explanatory Variable (Tk/year) Farm Non-farm Total Gender of household head ­0.0420 0.558** ­0.849** 0.040 (1 = male, 0 = female) Age of household head (years) 0.0002 ­0.002 0.011** 0.001 Maximum education of adult 0.0140** ­0.026** 0.140** 0.023** males (years) (continued ) Restoring Balance: Bangladesh's Rural Energy Realities 129 Table A2.3. Effects of Energy Use on Household Outcomes by Energy Source (N = 2,388) (Continued) Consumption Household Income (Tk/year) Explanatory Variable (Tk/year) Farm Non-farm Total Maximum education of adult ­0.0003 ­0.021* 0.073** 0.023** females (years) Log of household landholding 0.0480** 0.303** ­0.066 0.057** (decimal)1 Log of non-land household 0.0110** 0.167** 0.075** 0.048** asset (Tk) Energy use (kgoe/year) Biomass 0.0840** 0.528** ­0.323 0.144** Kerosene ­0.0160 0.232* ­0.182 ­0.037 Electricity 0.0520** ­0.025 0.042** 0.079** Proportion irrigated area 0.0080 ­0.151 ­0.091 0.034 in village Whether village has any Paved road ­0.0050 0.022 0.083 0.088** Primary school 0.0350 0.179* ­0.142 0.067* Health center ­0.0270 ­0.458** 0.424* ­0.068 Market ­0.0120 0.097 0.116 ­0.160 Commercial bank ­0.1020 ­0.010 ­0.060 ­0.220** NGO ­0.0760 0.036 0.062 0.095 Village price Rice (Tk/kg) ­0.0040 ­0.168** 0.159* 0.013 Flour (Tk/kg) ­0.0040 ­0.095** ­0.067 0.010 Potato (Tk/kg) 0.0110 ­0.011 ­0.010 ­0.008 Lentil (Tk/kg) ­0.0040 0.045** ­0.082** ­0.011** Fish (Tk/kg) 0.0010* ­0.004** ­0.0030 0.0010 Beef (Tk/kg) 0.0020 ­0.007 0.0460** 0.0060** Chicken (Tk/kg) ­0.0040** ­0.004 ­0.0005 ­0.0040** Soybean oil (Tk/liter) 0.0004 0.015 0.0370 ­0.0070 Milk (Tk/liter) 0.0080** 0.015 0.0010 ­0.0080 Sugar (Tk/kg) 0.0120* 0.027 ­0.0460 ­0.0050 Salt (Tk/kg) 0.0050 0.016 ­0.0930 0.0090 Village wage (Tk/day) Male 0.0030** ­0.005** 0.0170** 0.0040** Female ­0.0020** ­0.002 ­0.0110 ­0.0030* Child 0.0001 ­0.003 ­0.0020 0.0004 Region2 Dhaka ­0.0720* ­0.358** 1.2820** 0.1000* Chittagong ­0.2000** 0.200 0.2330 0.1670** Khulna ­0.1880** ­0.484** 1.2200** ­0.0460 Constant 5.2500** 2.142* 7.8920** 6.1630** R-squared 0.2100 0.327 0.0560 0.4310 Source: BIDS Survey (2004). Note: ** and * represent significance levels of 5 percent (or higher) and 10 percent, respectively. Consumption and income variables are in log form. 1 1 decimal = 1/100 acre. 2 Among the regions, Rajshahi is the excluded category. 130 World Bank Working Paper Chapter 5 Table A2.4. Tobit Estimate of Enterprise Energy Demand (N = 2,290) Energy Demand (Tk/month) Explanatory Variable Firewood Kerosene Diesel Electricity Age of enterprise (years) 0.156 0.967 ­0.215 0.354 Log of land asset (decimal)1 ­3.889** 0.925 ­1.062* 2.815 Log of non-land asset 8.164** 2.331 12.522** 52.211** (Tk 10,000) Enterprise type2 Agriculture and food 23.649** 88.441** ­25.595** 79.723** Manufacturing 27.822** ­123.354** 9.950* 266.723** Trade ­89.473** ­163.394** ­6.212 ­25.739 If village/growth center has 10.471 ­26.907 ­18.193** 418.724** electricity If village/growth center has ­1.015 ­39.292* ­3.127 63.939 paved road Price of energy source Fuelwood (Tk/kg) 5.246 1.168 0.932 ­10.455 Kerosene (Tk/liter) 11.438* ­1.028 ­3.758 ­14.563 Diesel (Tk/liter) ­4.788 ­11.166 3.287 23.379* Electricity (Tk/kWh) ­6.065 ­15.865 ­13.262* 25.676 Enterprise location3 Growth center ­5.148 48.409 2.351 205.446** Village ­13.800 58.144 ­2.199 198.329** Region4 Dhaka 10.324 1.224 ­5.623 118.407** Chittagong 9.598 97.450** ­21.479** 125.175** Rajshahi 13.801 ­6.282 0.036 34.074 Constant ­277.362* 317.274 ­84.736 ­1,455.246** Log likelihood ­1,689.142 ­11,097.043 ­687.906 ­13,539.475 Source: Data International (2002). Note: ** and * represent significance levels of 5 percent (or higher) and 10 percent, respectively. 1 1 decimal = 1/100 acre. 2 Service-based enterprises are excluded. 3 Home-based locations are excluded. 4 Khulna is excluded. The demand for enterprise energy is influenced more by growth-center/village character- istics than by the characteristics of individual enterprises; thus, a fixed-effects model that controls for growth-center or village-level heterogeneity was used for enterprise outcome regressions (see Tables A2.5 and A2.6). Restoring Balance: Bangladesh's Rural Energy Realities 131 Table A2.5. Effects of Enterprise's Total Energy Use on Profitability (N = 2,290) Explanatory Variable Revenue (Tk/year) Profit (Tk/year) Age of enterprise (years) 1,210.35 196.58 Log of land asset (decimal)1 11,467.57 1,516.09 Log of non-land asset (Tk) 95,605.74** 10,241.58** Enterprise type2 Agriculture and food 108,172.43** 52,372.62** Manufacturing 165,259.12 14,183.76 Trade 236,121.32** 8,902.59* Total energy use (Tk/year) 0.86 0.55** Constant ­410,412.62 60,624.86* F-statistics (9,2061) 11.28 12.42 Source: Data International (2002). Note: ** and * represent significance levels of 5 percent (or higher) and 10 percent, respectively. 1 1 decimal = 1/100 acre. 2 Service-based enterprises are excluded. Table A2.6. Effects of Enterprise's Energy Use on Profitability by Source (N = 2,290) Explanatory Variable Revenue (Tk/year) Profit (Tk/year) Age of enterprise (years) 990.158 153.033 Log of land asset (decimal)1 11,273.32 1,424.647 Log of non-land asset (Tk) 95,669.08** 10,178.643** Enterprise type2 Agriculture and food 106,775.51** 49,891.651** Manufacturing 119,636.91 5,690.698 Trade 229,058.11** 7,567.878 Energy use (Tk/year) Biomass 5.716* 1.217** Kerosene/diesel ­3.872 ­0.271 Electricity 4.107 1.351** Constant ­403,568.82 62,085.22 F-statistics (11,2059) 9.55 11.68 Source: Data International (2002). Note: ** and * represent significance levels of 5 percent (or higher) and 10 percent, respectively. 1 1 decimal = 1/100 acre. 2 Service-based enterprises are excluded. ANNEX 3 Consumer's Surplus as a Measure of Welfare C onsumer's surplus--a measure of well being that relies on the difference between what a person, household, or group is willing to pay for energy and what must be paid--has a long history in economics as a method for estimating the benefits of public projects.1 The first use of applying consumer's surplus for valuing the benefits of electricity in the World Bank was seen as early as 1975.2 This was followed by a more elab- orate exploration of the concept in the mid-1980s.3 More recently, the increasingly pop- ular approach is being applied in most rural electrification projects, including such countries as Bolivia, Lao PDR, Peru, and the Philippines.4 Although fairly easy to apply, the procedure is not always well understood and is not without its critics, even by those educated in its underlying principles. The following sec- tions provide a brief primer on the concept and address some of the more frequent criti- cisms that emerge when applying this method. One criticism not addressed is that money and well being are not necessarily the same thing; it therefore questions whether any mon- etary measure of benefit is valid. Consumer Demand The starting point for an understanding of consumer's surplus is the consumer demand curve (see Figure A3.1). The demand curve shows the relationship between the price facing the consumer and the quantity consumed at that price. Of course, the quantity of energy consumed at any This section is adapted from Henry M. Peskin, "A Primer on Consumer Surplus and Demand: Common Questions and Answers," ESMAP Knowledge Exchange Series, No. 5, May 2006. 133 134 World Bank Working Paper point in time depends on far more Figure A3.1. Consumer Demand Curve than its price--the weather, the taste for energy-consuming items such as radio and TV, the need to support energy needs of business, and most importantly, the consumer's income. Clearly, if one chooses to focus only on the relationship between energy prices and energy consumed, these other factors must somehow be held constant. In general, if all the non- price factors remain fixed, the higher the price of a good (in this case Source: Peskin (2006); and World Bank (2002c). energy), the less likely the consumer will demand it. The demand curve shows price­quantity relationship for an individual, a household, a group of individuals, or a group of households. The interpretation of the curve differs if it shows an individual or a group, especially if the group contains a mix of incomes and tastes. For a heterogeneous group, the tendency for higher prices leading to lower energy con- sumption usually implies that most of the consumption at high prices is by those individ- uals with higher incomes or with a stronger desire for energy. For an individual, the lower consumption at higher prices usually means that the individual will tend to substitute for other goods that provide a less costly means to maximize his or her satisfaction or well being. In either case, if the market price of energy is below what some people would be will- ing to pay for it, these people would experience a gain in their well being. This gain is the principal argument for using consumer's surplus as a benefit measure (see Figure A3.2). If the price of energy fell from P0 to P1, those individuals (or a particular individual) who would be willing to pay P0 in order to consume Q0 of energy, can now consume the Q1 of energy at the lower price. The original "Q0 consumers" (or in the case of an individual, the "Q0 consumption") realize a benefit of P0 ­ P1 for each of the units of Q0 consumed. This benefit would be realized for any consumption less than Q1 corre- sponding to any price higher than P1, Figure A3.2. Consumer's Surplus such as for those willing to pay P* for Q* units of energy. Thus, when one considers all consumption­price combinations between the original Q0, P0 combination and the new Q1, P1 combinations, the total of all gains is represented by the shaded triangle a,b,c plus the rectangular area P1, P0, a, b. In brief, energy is demanded for the service that it provides. There- fore, to calculate consumer's surplus requires an estimate of demand for Source: Peskin (2006); and World Bank (2002c). Restoring Balance: Bangladesh's Rural Energy Realities 135 lighting, entertainment, communications, or other services closely linked to energy. For lighting, the demand is quantified as kilolumen hours and for entertainment, it is radio or television listening or viewing hours. Thus, it is possible to obtain a measure of con- sumer's surplus by using the price and quantity of kilolumen or radio listening hours for households using kerosene, batteries, or electricity from a grid system.5 This annex assumes that the benefit measured by consumer's surplus is a satisfactory measure of the benefit of policy that brings about lower energy prices and considers key challenges in this approach. Estimating the Demand Curve Consumer's surplus greatly depends on the shape of the demand curve, particularly between the original Q0, P0 combination and the new Q1, P1 combination. If the curve were not a sim- ple straight line but instead bowed or bent toward the origin of the graph, such as indicated by the dotted line, the consumer's surplus would be smaller, as shown by the more lightly shaded area. Although the consumer's surplus could be far smaller than that measured in the World Bank's ESMAP energy studies, the investigators chose not to assume that it is smaller.6 Instead, the studies based the demand curve (and the resulting measure of consumer's surplus) on actual price­quantity observations drawn from household surveys. For com- binations not observed, there was a simple linear extrapolation between points that were observed. Of course, the extrapolated points may not fall on the "true" (but unobservable) demand curve. However, since the extrapolations rely solely on what can be observed, the investiga- tors believe that this is a more honest approach than simply assuming some sort of curva- ture between the observable points. Such assumptions are necessarily arbitrary and could produce a wide range of results. Besides the problem of estimating the correct demand curve, application of the consumer's surplus method has raised a number of questions and criticisms.7 Notes 1. See Alfred Marshall, Principles of Economics, 8th ed., MacMillan and Co., London, 1930; Hal R. Varian, Microeconomic Analysis, 3rd ed., W. W. Norton and Co., New York, 1978. 2. See Dennis Anderson, Cost and Benefits of Rural Electrification: A Case Study of Costa Rica, World Bank Public Utilities Report No. RES 5, World Bank, Washington, DC, 1975. 3. See David Pearce and Michael Webb, Economic Benefits of Electricity Supply, Energy Department Paper No. 25, World Bank, Washington, DC, 1985. 4. See Kyran O'Sullivan and Douglas F. Barnes, Energy Policies and Multitopic Household Surveys: Guidelines for Questionnaire Design in Living Standards Measurement Study, Energy and Mining Sector Board, Discussion Paper No. 17, World Bank, Washington, DC, 2006. 5. See World Bank, Rural Electrification in the Philippines: Measuring the Social and Economic Benefits, ESMAP Report 225/02, Washington, DC, 2002. 6. Ibid. 7. Peskin, 2­4. References Asaduzzaman, M., and A. Latif. 2005. "Energy for Rural Households: Towards a Rural Energy Strategy in Bangladesh." Bangladesh Institute of Development Studies, Dhaka. Barkat, A., M. Rahman, S. Zaman, A. Podder, S. Halim, N. Ratna, M. Majid, A. Maksud, A. Karim, and S. Islam. 2002. "Economic and Social Impact Evaluation Study of the Rural Electrification Program in Bangladesh." Report to National Rural Electric Cooperative Association (NRECA) International, Dhaka. BBS (Bangladesh Bureau of Statistics). 2005. Report of the Agricultural Sample Survey of Bangladesh, vol. 1. Dhaka: Bangladesh Bureau of Statistics. Dasgupta, S., M. Huq, M. Khaliquzzaman, K. Pandey, and D. Wheeler. 2004. Who Suffers from Indoor Air Pollution?: Evidence from Bangladesh. World Bank Policy Research Working Paper 3428. Washington, DC: World Bank. Data International. 2004. "Energy Usage by Rural Enterprises of Bangladesh." Background Report, Dhaka. Ghimire, Prakash C. 2005. "Final Report on Technical Study of Biogas Plants Installed in Bangladesh." Report submitted to National Programme on Domestic Biogas in Bangladesh (partnership of Netherlands Development Organization and IDCOL), Dhaka. GOB (Government of the People's Republic of Bangladesh). 1987. Bangladesh Energy Plan- ning Project: Final Report, vols. I­VII. Prepared by Sir Halcrow and Partners, Motor Columbus Consulting Engineering, Inc., Petronus Consultants in Association with TSL and PSL, Dhaka. ------. 1996. "National Energy Policy," Ministry of Energy and Mineral Resources. Bangladesh Gazette (January), Dhaka. ------. 2005. "Unlocking the Potential: National Strategy for Accelerated Poverty Reduc- tion." Dhaka: General Economics Division, Planning Commission. 137 138 World Bank Working Paper Hossain, M. M. Golam. 2003. "Improved Cookstove and Biogas Programs in Bangladesh." Energy for Sustainable Development 7(2): 97­100. Hutton, G., and E. Rehfuess. 2006. "Guidelines for Conducting Cost-Benefit Analysis of Household Energy and Health Interventions to Improve Health." Paper prepared for World Health Organization, Geneva. Hutton, G., E. Rehfuess, F. Tediosi, and S. Weiss. 2006. "Evaluation of the Costs and Benefits of Household Energy and Health Interventions at Global and Regional Levels." Paper prepared for World Health Organization, Geneva. IEA (International Energy Agency). 2003. Key World Energy Statistics 2003. Paris. Islam, M. N. 1980. "Village Resources Survey for the Assessment of Alternative Energy Technology." Report submitted to International Development Research Centre, Bangladesh University of Engineering and Technology, Dhaka. ------. 1986. Rural Energy Survey in the Third World: A Critical Review of Issues and Methods. Ottawa: International Development Research Centre. ------. 2001. "Energy Strategy for Rural Bangladesh." Background paper prepared for World Bank, Dhaka. Khalequzzaman, M. 2005. "Assessment of Rural Energy Delivery Mechanisms in Bangladesh." Background study prepared for World Bank Energy Sector Manage- ment Assistance Program (ESMAP), Washington, DC. Kulkarni, V., Douglas Barnes, and Sandro Parodi. 2007. "Rural Electrification and School Attendance in Nicaragua and Peru." Draft paper, World Bank, Washington, DC. Leach, Gerald. 1987. "Household Energy in South Asia." Biomass 12: 155­184. McKenzie, Wood. 2006. "Gas Sector Master Plan and Strategy for Bangladesh." Interim report prepared for Petrobangla, Dhaka. Nexant. 2006. "Power Development Program, Draft Final Report: Master Plan Update." Prepared for Asian Development Bank and Ministry of Power, Energy, and Mineral Resources. Openshaw, Keith. 2004. "Bangladesh: Biomass Energy Supply." Background study pre- pared for World Bank Energy Sector Management Assistance Program (ESMAP), Washington, DC. O'Sullivan, Kyran, and Douglas F. Barnes. 2006. Energy Policies and Multitopic House- hold Surveys: Guidelines for Questionnaire Design in Living Standards Measurement Studies. Energy and Mining Sector Board, Discussion Paper No. 17. Washington, DC: World Bank. Pitt, Mark M., Mark R. Rosenzweig, and Md. Nazmul Hassan. 2005. Sharing the Burden of Disease: Gender, the Household Division of Labour and the Health Effects of Indoor Air Pollution. CID Working Paper No. 1119. Boston: Center for International Devel- opment, Harvard University. United Nations. 2005. Energy Services for the Millennium Development Goals. Joint publication of UN Millennium Project, United Nations Development Programme (UNDP), World Bank, and Energy Sector Management Assistance Program (ESMAP). Washington, DC and New York: World Bank/ESMAP and UNDP. World Bank. 2002a. Energy Strategy for Rural India: Evidence from Six States. ESMAP Formal Report 258/02. Washington, DC. ------. 2002b. India: Household Energy, Indoor Air Pollution, and Health. ESMAP Formal Report 261/02. Washington, DC. ------. 2006. World Development Report 2007: Development and the Next Generation. Washington, DC. WHO (World Health Organization). 2006. Fuel for Life: Household Energy and Health. Geneva. Glossary aman rice planted in July­September and harvested in November­December aus rainfed summer rice planted in March­April biri tobacco boro irrigated winter rice planted in December­February and harvested in April­June gur molasses hachak (see petromax) kupi/cherag traditional kerosene lamp motive power energy source used for powering machinery paan betel quid petromax pressurized kerosene lamp thana/upazila rural subdistrict 139 Special Report Series Renewable Energy Thematic Area East Asia and Pacific Region (EAP) Sustainable and Efficient Energy Use to Alleviate Indoor Air Pollution in Poor Rural Areas in China (002/07) Latin America and the Caribbean Region (LCR) Nicaragua Policy & Strategy for the Promotion of Renewable Energy Resources (003/07) Global (GLB) Considering Trade Policies for Liquid Biofuels (004/07) Energy Poverty Thematic Area Global (GLB) Risk Assessment Methods for Power Utility Planning (001/07) South Asia Region (SAR) Restoring Balance: Bangladesh's Rural Energy Realities (006/09) Energy Security Thematic Area Global (GLB) Coping with High Oil Price Volatility (005/08) 141 Eco-Audit Environmental Benefits Statement The World Bank is committed to preserving Endangered Forests and natural resources. We print World Bank Working Papers and Country Studies on 100 percent postconsumer recy- cled paper, processed chlorine free. The World Bank has formally agreed to follow the rec- ommended standards for paper usage set by Green Press Initiative--a nonprofit program supporting publishers in using fiber that is not sourced from Endangered Forests. For more information, visit www.greenpressinitiative.org. In 2007, the printing of these books on recycled paper saved the following: Trees* Solid Waste Water Net Greenhouse Gases Total Energy 264 12,419 96,126 23,289 184 mil. * 40" in height and 6-8" in diameter Pounds Gallons Pounds CO2 Equivalent BTUs Restoring Balance: Bangladesh's Rural Energy Realities is part of the World Bank Working Paper series. These papers are published to communicate the results of the Bank's ongoing research and to stim- ulate public discussion. Just less than one-half of the people in developing countries have no access to electricity, and a similar number are reliant on biomass en- ergy for cooking and heating. As a consequence, they are deprived of the means of moving out of poverty. Greater access to modern en- ergy services can improve poor people's income through enhance- ment of productive use of energy and it can also increase their quality of life by providing quality lighting, communication, and other im- portant services. The Energy Sector Management Assistance Program (ESMAP) is a global technical assistance program administered by the World Bank that assists low- and middle-income countries to acquire know-how and to increase institutional capability to secure clean, reliable, and affordable energy services for sustainable economic development. ESMAP was established in 1983 under the joint sponsorship of the World Bank and the United Nations Development Programme as a partnership in response to global energy crises. Since its creation, ESMAP has operated in over 100 different coun- tries through more than 500 activities covering a broad range of en- ergy issues. World Bank Working Papers are available individually or on standing order. Also available online through the World Bank e-Library (www.worldbank.org/elibrary). ISBN 978-0-8213-7897-7 THE WORLD BANK 1818 H Street, NW Washington, DC 20433 USA Telephone: 202 473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org SKU 17897