W O R L D B A N K W O R K I N G P A P E R N O . 2 1 Revisiting Reform in the Energy Sector Lessons from Georgia Julian A. Lampietti Hernan Gonzalez Edition Margaret Wilson Ellen Hamilton Sergo Vashakmadze Bilingual 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 . 2 1 Revisiting Reform in the Energy Sector Lessons from Georgia Julian A. Lampietti Hernan Gonzalez Margaret Wilson Ellen Hamilton Sergo Vashakmadze THE WORLD BANK Washington, D.C. Copyright © 2004 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: December 2003 printed on recycled paper 1 2 3 4 05 04 03 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 typescript of this paper therefore has not been prepared in accordance with the procedures appropriate to journal printed texts, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank cannot 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 on the part of the World Bank any judgment of the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. The World Bank encourages dissemination of its work and normally will grant permission for use. Permission to photocopy items for internal or personal use, for the internal or personal use of specific clients, or for educational classroom use, is granted by the World Bank, provided that the appropriate fee is paid. Please contact the Copyright Clearance Center before photocopying items. Copyright Clearance Center, Inc. 222 Rosewood Drive Danvers, MA 01923, U.S.A. Tel: 978-750-8400 · Fax: 978-750-4470. For permission to reprint individual articles or chapters, please fax your request with complete information to the Republication Department, Copyright Clearance Center, fax 978-750-4470. All other queries on rights and licenses should be addressed to the World Bank at the address above, or faxed to 202-522-2422. ISBN: 0-8213-5689-5 eISBN: 0-8213-5690-9 ISSN: 1726-5878 Julian Lampietti is Senior Social Development Economist in the Environmentally and Socially Sus- tainable Development Sector Unit of the Europe and Central Asia Region at the World Bank. Hernan Gonzalez is Consultant to the World Bank. Margaret Wilson is Consultant to the World Bank. Ellen Hamilton is Urban Planner in the Europe and Central Asia Infrastructure and Energy Services Department of the World Bank. Sergo Vashakmadze is Economist in the Europe and Central Asia Poverty Reduction and Development Management Department of the World Bank. Cover Photo: Replacing the electricity meter in Tbilisi, Georgia. Source: AES-Telasi. Library of Congress Cataloging-in-Publication Data has been requested. TABLE OF CONTENTS Foreword v Abstract vii Preface ix Acronyms and Abbreviations xi Executive Summary 1 1. Introduction 3 Data and Research Design 4 2. Energy Sector Reform Context 5 The Economy and Household Welfare 5 Access to Network Energy 6 Power Sector Reform 7 Energy Tariffs 8 3. Impact on Households 11 Regional Differences in Energy Expenditures 11 Composition of Energy Expenditures 12 Changes in Electricity Consumption--Tbilisi 14 Health and the Environment 15 Conclusions 17 4. Impacts on Utilities 19 Performance of AES Telasi 20 Prices 20 Subsidies 21 Service Quality 21 Enforcement 22 Structure of Arrears 23 Conclusions 23 5. Impacts on Government 25 State Support for the Energy Sector 25 Municipal Support for the Energy Sector 26 Electricity Subsidy Effectiveness 27 Conclusions 30 Annexes 31 Annex A. Converting Energy Prices into Cost per Effective BTU 31 Annex B. Environmental Outcomes 33 Annex C. Analysis of Telasi's Revenues 35 LIST OF FIGURES Figure 1. Stated and Actual Household Electricity Payments 4 Figure 2. Household Expenditure Shares by Quarter 6 Figure 3. Power Sector Reform Milestones in Georgia 7 Figure 4. Effective Energy Prices, GEL per Million BTU 9 iii Figure 5. Share of Utilities in Total Expenditure by Region 12 Figure 6. Total Household Energy Consumption in Effective BTUs--Tbilisi 13 Figure 7. Energy Expenditure Shares by Fuel Type--Tbilisi 13 Figure 8. Household Electricity Consumption--Tbilisi 14 Figure 9. Distribution of Electricity Consumption--Tbilisi, 2002 15 Figure 10. Energy Expenditure Shares on Clean and Dirty Fuels--Tbilisi 16 Figure 11. Demand for Fuelwood in Georgia (winter 2002) 16 Figure 12. Collection Rates by Re-metered Status--Tbilisi 22 Figure 13. Collection Rates by Quintile--Tbilisi 23 Figure 14. Frequency of Household Electricity Consumption (kWh per year)--Tbilisi 29 LIST OF TABLES Table 1. Aggregate Impact of Reform on Collection Rates--Tbilisi 21 Table 2. State Budget Payments to the Energy Sector 2001-2003 (thousand GEL) 26 Table 3. State Budget Energy Subsidies--Tbilisi (thousand GEL) 27 Table 4. Electricity Subsidy Incidence--Tbilisi 28 Table 5. Subsidy Coverage--Tbilisi 28 Table 6. Simulation of Subsidy Cost-Effectiveness--Tbilisi 29 Table A.1. Calculation of Cost per Effective BTU 32 Table C.1. Average Hours of Electricity and Amount Paid by Region 36 Table C.2. Regression Model 37 Table C.3. Regression Results 37 Table C.4. Effect of Collection Rate on Revenues 38 Table C.5. Effect of Enforcement on Revenues 39 iv FOREWORD E nergy issues remain intensely problematic for the energy-poor countries in the former Soviet Union as this case study of Georgia, one of the better reformers, shows. In the early transi- tion years, Georgian citizens lost energy for light, cooking and heat in parallel with a general col- lapse in GDP and increase in poverty. In response, the Government embarked on a series of reforms primarily intended to improve the troubled electricity sector. Generation and supply were unbundled to introduce competition, a regulatory framework was introduced, and the Tbilisi elec- tricity distribution company was privatized to an American investor in the first such case in the former Soviet Union. The scope of electricity reform in and of itself would make a thought provoking case study; however, this paper reviews the effects of reform in both the electricity and gas sectors, as well as the related interactions between the two sectors. Both are examined from the perspective of households, utility operators, and the government in order to highlight lessons from the reform experience. The study includes analysis of expenditures on electricity, accumulated arrears, and subsidy cost-effectiveness by different income groups over the last five years. This allows us to assess the impacts of energy sector reform on these primary stakeholders: households, utilities, and the government. The results of the study should be helpful for further implementation of utility reforms in Georgia and should supply the Government, the World Bank, other donors and other countries facing similar problems with a thoughtful analysis of the many aspects of the Georgian experience to drawn on for future programs. Laura Tuck Sector Director Europe and Central Asia Region v ABSTRACT T his paper reviews the changes in the supply of electricity and gas from the perspective of households, utility operators, and the government. The objective is to highlight lessons from the reforms implemented and to apply them to the future reform program planned for the rest of the energy sector. The paper concludes that improved service quality and the increased supply of clean and subsidized natural gas have offset the potentially negative impact of higher electricity prices. Despite very good performance by the privatized electricity distribution company in Tbilisi, the sustainability of the reform program is still in doubt. Consolidated government expenditures on energy have increased, but to a large extent this simply recognizes costs that were incurred, but not paid, prior to reform. Existing subsidies to households for electricity provide compensation beyond levels that produce large welfare gains. Changing the subsidy system to base targeting on actual levels of electricity consumption while providing enough compensation to ensure the household received a basic level of electricity, would be one option to improve subsidy targeting. vii PREFACE T his paper is based on research carried out between October 2002 and June 2003 under the sponsorship of Donna Dowsett-Coirolo (Country Director, ECCU3). Additional funding was provided by the Norwegian Environmental Trust Fund. The research team included both Geor- gian and international specialists. A Working Group, consisting of representatives of Georgian gov- ernment agencies and NGOs, provided commentary and input at several stages throughout the study. Members of this group included Akaki Zoidze, Natia Turnava, Nodar Kapanadze, Ignacio Iribarren, Irakli Avaliani, David Gzirishvili, and Devi Khechinashvili. Special acknowledgement should be given to USAID, IMF, the Georgia State Department of Statistics, to AES Telasi, and to Save the Children--all of which provided full access to their detailed databases and extensive support in interpreting the results. The initial findings were presented to the World Bank in Washington, D.C. in May 2003 and to the Working Group in June 2003. The report was revised following feedback from these two sources. Julian Lampietti (ECSSD) was the primary author of the report, with support from Ellen Hamilton (ECSIE), Hernan Gonzalez (consultant), Margaret Wilson (consultant), Sergo Vashka- madze (ECSPE), and Taras Pushak (consultant). Peer reviewers were Kirk Hamilton (ENV) and Anis Dani (SDV). Helpful comments were provided by Brian Smith (ECSIE), Bjorn Hamso (ECSIE), Aleksandra Posarac (ECSHD), Wojciech Maliszewski (IMF), and Rocio Castro (ECSPE). ix ACRONYMS AND ABBREVIATIONS ARI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Acute Respiratory Infections GDP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Gross Domestic Product GEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Georgian Lari GNERC . . . . . . . . . . . . . . . . . . . .Georgian National Energy Regulatory Commission GUDC . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Georgian United Distribution Company GWEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Georgian Wholesale Electricity Market HBS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Household Budget Survey kWh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Kilowatt Hour LPG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Liquid Propane Gas STC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Save the Children VAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Value Added Tax WHAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Winter Heat Assistance Program Exchange rate in 2002: US $1 = GEL 2.15 xi EXECUTIVE SUMMARY T his paper reviews recent changes in Georgia in the supply of electricity and gas--examining them from the perspective of households, utility operators, and the government. The objec- tive is to highlight the lessons from reforms implemented fully or partially and to apply them to the two sectors. With living standards falling over the last decade, the government has been reforming the power sector, partially privatizing it, and establishing a legal and regulatory framework to govern it. Households continue to have high levels of access to network services, but the numbers mask poor service and supply shortages, particularly outside Tbilisi. Despite rapid increases in the price of electricity and non-network fuels, the share of spending on energy has remained constant. In Tbilisi, improved service quality and increased supply of clean and inexpensive natural gas appear to have offset the burden of higher electricity prices. The shift to natural gas also means that dirty fuel consumption may not have had big negative externalities. Outside Tbilisi, however, electricity price increases have not been offset by better service and the access to gas has not increased. Many households continue to burn wood, especially in rural areas. So moves to wood- burning technology that reduce the cost per BTU could produce important welfare gains. Households in Tbilisi are consuming about 125 kWh per month of electricity--close to basic minimum needs. Demand appears to be fairly inelastic, suggesting large welfare losses with future price increases (as well as large revenue gains to the utility). A kinked demand curve suggests that the welfare gains from providing households with large electricity subsidies (more than 150 kWh per month) are probably small. Careful consideration must go to the welfare effects of future price increases and to the design of the most appropriate mitigation measures. Despite very good performance by the main private operator, the sustainability of the reform program is still in doubt. Over the last three years AES Telasi has increased its receipts by 135 per- cent. The data suggest that re-metering is as important as price as a determinant of utility receipts; and it may even be more important in the early stages of reform. The data also suggest that an 1 aggressive approach to reducing nonpayment does not necessarily have a disproportionate adverse impact on low income households--particularly if suitable subsidy and transfer mechanisms are in place. Government expenditures on the energy sector, as recorded in the consolidated budget, have increased since the implementation of sector reforms. Many of these expenditures are simply the recognition of costs that were incurred but not paid prior to reform. Government subsidies to the sector are also growing. This is due to increasing tariffs and to government decisions to increase support for specific programs. Of particular concern is the sustainability of gas subsidies (provided by both the state and municipal budgets) as additional households are connected to the gas supply network while cost recovery measures are not in place. Without access to clean and inexpensive natural gas to offset the potential negative impact of higher electricity prices there may be a shift towards more dirty fuel consumption, particularly by the poor. If subsidies are to be a tool of poverty alleviation, the merits of the current system are dubi- ous. A significant part of subsidies goes to households in the higher expenditure quintiles. In addition, a large share of the subsidies--at least under the expanded program of privileges--is compensation for electricity consumption beyond levels that produce large welfare gains. Re-orienting the subsidy program to maximize the compensation for welfare losses would benefit both the consumers and the government budget. One way to do this would be to base targeting on actual levels of electricity consumption--and to provide enough compensation to ensure that the household receives a basic level of electricity. Such a program would provide a rel- atively simple mechanism for targeting and be more cost-effective in terms of the welfare gain per unit of subsidy paid. This new subsidy program could be piloted as part of the new management changes being put into place for the Georgian United Distribution Company (GUDC). To monitor the poverty tar- geting of the subsidy, the Household Budget Survey (HBS) could be linked directly to the utili- ties' billing and payment database. Over time--as data on consumption patterns, income, and pay- ment are collected and analyzed--the targeting system could be refined and the overall cost reduced. CHAPTER 1 INTRODUCTION O ne of the harsher realities of independence for the former Soviet republics has been the loss of subsidized transfers from the center for fuel and utilities. In the years since independence, Georgians, with other "energy poor" republics, have been subject to higher costs and declining service levels for household utilities--particularly energy. The combination of low household incomes, high international prices for fuel, the need for utilities to rely on internally generated funds for capital investment, and the political ramifications of removing subsidies at a time of gen- eral economic decline have led to a "worst of all worlds" situation. In the light of these problems, several countries, Georgia among them, concluded that state own- ership and management of utilities was not sustainable. The investment capital and efficiency improve- ments needed for the utilities could be better achieved through increased competition and private sec- tor participation, particularly in the supply of energy services. This might, at least in the short term, lead to higher utility prices, which could have an adverse impact on the low-income households. In the expectation that gains in efficiency and service quality would, over time, offset welfare losses from higher prices and potential negative externalities from restricted access, the govern- ment of Georgia--with the support of the donor community--undertook a program of utility sec- tor reform. Not all the reforms were concurrent, and some have not yet begun. But six years have passed since the start of the process, making it appropriate to reflect at this point on the results of the program to date. This paper reviews the changes in the supply of electricity and gas from the perspective of households, utility operators, and the government. The objective is to highlight lessons from the reforms implemented (fully or partially) and to apply them to the future reform program planned for the rest of the energy sector. The first chapter provides background and context material for the study. It describes recent economic and poverty trends, as well as main features of the energy sector reform program. The following three chapters outline the effects of the energy reform pro- gram on households, utilities, and the government. 3 4 WORLD BANK WORKING PAPER Data and Research Design The information in the report is drawn from government statistical databases, surveys by NGOs, statistical information from service providers, interviews with government, utilities and NGOs, and focus group sessions with sector experts and household representatives. The statistical analysis of the household and utility impacts of reform is based primarily on data from three sources: the Household Budget Survey (HBS), carried out quarterly since 1996 by the Georgia State Department of Statistics1; the Multi-Sector National Survey of Households in Georgia 2002,2 carried out in February 2002 by Save the Children (STC); and the electricity consumption, billing, and payment data from AES Telasi for households in Tbilisi in the HBS from 2000 to 2002. An important feature of the study was the ability to merge data sets from the HBS and AES Telasi to link such household characteristics as income to household electricity con- sumption and payment patterns. Merging the HBS and AES Telasi data sets revealed important discrepancies in reported elec- tricity payments. The HBS data are based on households' self-reported electricity payments--the AES Telasi data, on household payments recorded in the customer's billing and payment records. A comparison of corresponding data (for the same household in the same month) revealed that payments reported in the HBS were consistently higher than those recorded by Telasi in 2000 and 2001 (figure 1). This may be due to corruption3 or recall error (households might report bills rather than payments). Despite these differences, the data sets provide a sound basis for the analy- sis because both follow the same increasing trend in payments and the difference between the two narrows over time. FIGURE 1: STATED AND ACTUAL HOUSEHOLD ELECTRICITY PAYMENTS 20 16 Stated payment 12 (HBS) month per 8 GEL Actual payment 4 (AES Telasi) 0 Jan-00 May-00 Sep-00 Jan-01 May-01 Sep-01 Jan-02 May-02 Sep-02 Source: Georgia Household Budget Survey, AES Telasi. 1. State Department for Statistics of Georgia, "Poverty Monitoring in Georgia: Annual Report 2000," Tbilisi, 2001. 2. This survey was funded by USAID. The authors are Larry Dershem and Irakli Sakandelidze. 3. Households paying more to meter readers than the meter readers transfer to the utility. Focus group sessions suggested that this was a serious problem in the past, although the incidence has decreased with the installation of new meters and better control by Telasi. CHAPTER 2 ENERGY SECTOR REFORM CONTEXT The Economy and Household Welfare Beset with civil war, the loss of markets, and the loss of low cost resources, Georgia's GDP fell by 70 percent from 1990 to 1994. The situation has stabilized in recent years, but recovery has been slow. GDP, having reached $3.2 billion in 2001, is estimated to have grown a further 5.4 percent in 2002. But as the recent Poverty Assessment4 notes, GDP growth has not translated into corresponding improvements in living standards because it has been weak and concentrated in too few sectors. The Poverty Assessment concluded that poverty increased steadily in recent years.5 Since 1996 average consumption has fallen, inequality has risen, and living standards have declined, with households forced to shift to a lower quality basket of goods and services. In real terms average monthly per capita expenditure fell 4 percent from late 1996 to late 2002, dropping 18 percent to the end of 2001 before recovering in 2002. A large part of the increase between 2001 and 2002 was driven by an increase in monetary and nonmonetary food expenditures. Expenditures on household goods, personal items, and services (such as health and education) increased slightly. Overall, the food share of household expenditures decreased by approximately 2 percentage points between 1996 and 2002, while the share allocated to household and personal items remained rela- tively stable (figure 2). The share of expenditures on utilities (energy and water) has held more or 4. World Bank report No. 19348-GE, "Georgia Poverty and Income Distribution", May 1999 and World Bank report No. 22350-GE, "Georgia Poverty Update", January 2002. 5. This is not consistent with National Accounts data, which show an increase in household expenditures. A possible explanation for this apparent contradiction is given in World Bank report No. 22350-GE, "Georgia Poverty Update", January 2002, page 58, paragraph 35. See also Ravallion (2001), "Measuring Aggregate Welfare in Developing Countries: How well do National Accounts and Surveys Agree?" (WB Working Paper No.2665). 5 6 WORLD BANK WORKING PAPER FIGURE 2: HOUSEHOLD EXPENDITURE SHARES BY QUARTER 80% Food 60% Expenditures Other 40% expenditures Household 20% of Utilities Share 0% II IV III II IV III II IV III II IV III II IV III II IV III IV 1996-III 1997-I 1998-I 1999-I 2000-I 2001-I 2002-I Source: Georgia Household Budget Survey. less constant at about 8 percent, with seasonal variation up to 10 percent driven by winter heating requirements. The decline in household expenditures was not uniform across the country. In Tbilisi house- hold expenditures fell by 2 percent over 1996­2002. In other cities it dropped by 20 percent. And in rural areas consumption initially fell by 17 percent (1997­2001), but increased by 28 per- cent in 2002 for a net increase of 7 percent over the period. Access to Network Energy Access to network energy has also changed over time. District heating disappeared in the late 1990s. For electricity, 98 percent of households remain connected to the network. But supply has failed to meet demand owing to a combination of factors. Drought reduces the availability of hydroelectricity. External arrears have reduced the ability to import electricity from neighboring countries. And an explosion at the Gardabani thermal plant reduced thermal generation by half for much of the winter of 2001. In Tbilisi service has improved over the last few years, except for February and March of 2001, when less than half of total demand was supplied. Outside Tbilisi, however, supply constraints are severe and persistent, with households receiving 4.5 to 17 hours of electricity per day, depending on location.6 For natural gas, the number of connections increased in Tbilisi, particularly in 2001 and 2002.7,8 Outside the capital, however, the number of connected households has fallen ­ possibly 6. Save the Children, Multi-Sectoral National Survey of Households in Georgia 2002. 7. Tbilgazi's customer base increased from 39,000 households in June of 2000 to 164,000 households in January of 2003. There are approximately 300,000 households in Tbilisi. 8. In the HBS households were asked if they had a natural gas connection. These data indicate that the number of connections decreased nationwide from 1998 to 2000, with a small increase in Tbilisi from 2000 to 2001. REVISITING REFORM IN THE ENERGY SECTOR 7 due to limited or non-existent service. Gas supply has been intermittent but appears to be stabiliz- ing as external arrears are paid off.9 The completion of the Baku-Tbilisi-Ceyhan pipeline is expect- ed to further reduce supply constraints by providing an alternative to Russian gas imports. Power Sector Reform Power reform in Georgia generally followed the World Bank's 1998 Europe and Central Asia energy sector strategy (figure 3).10 This included unbundling to introduce competition in genera- tion and supply, establishing predictable and transparent regulations, selling assets to private strate- gic investors, and raising prices to cost recovery levels. The intended outcomes were lower fiscal deficits, more efficient use of resources, higher production efficiency, and better consumer service. In Georgia the vertically integrated electricity enterprise--Sakenergo--was divided into several generation enterprises, and separate transmission and dispatch companies (which were recently re- merged). Distribution was divided into regional companies, and management was devolved to local administrations. The same pattern occurred in the gas sector, though most of the gas was imported from Russia. Concurrent with the unbundling, an electricity law was drafted (and subsequently amended to cover gas supply), and an independent regulatory authority, the Georgia National Energy Regulatory Commission (GNERC) was established. GNERC was able to raise prices for electricity and natural gas to cost recovery levels (including full depreciation and a return on investment)-- although restrictions in the legislation for allowable claims related to nonpayment and nontechni- cal losses meant that the tariffs in many instances were insufficient to provide a satisfactory rev- enue flow. In 1999 a Wholesale Electricity Market (GWEM) was set up to manage the flow of payments among sector enterprises. FIGURE 3: POWER SECTOR REFORM MILESTONES IN GEORGIA 0.16 Regulatory agency Electricity tariff 0.12 established Sakenergo unbundled kWh 0.08 per Telasi privatized GEL 0.04 GWEM established UDC established 0 1997 1998 1999 2000 2001 2002 Source: Georgia National Energy Regulatory Commission (Annual Reports) and personal interviews. 9. Gas is purchased from the Russian company Itera by industrial customers, from the Gardabani power plant, and from the local gas distribution companies. In the past, Itera has tied gas delivery to payments from any and all of these customers. So, if one or more customers accumulated significant arrears, gas supply to the country was curtailed until a satisfactory settlement could be reached. 10. World Bank. "Energy in Europe and Central Asia: A Sector Strategy for the World Bank Group." World Bank Discussion Paper No. 393. World Bank. Washington, D.C. 1998. pp. 29-30. 8 WORLD BANK WORKING PAPER In 1998 the government invited tenders for several of these newly restructured enterprises, including the generation assets and the electricity and gas distribution companies.11 Gas distribu- tion companies in urban centers outside Tbilisi were sold to Sakgazi--a joint-venture between local partners and the Russian gas supply company, Itera. Telasi, the electricity distribution compa- ny serving Tbilisi, was sold to AES at the end of 1998, and some of the smaller electricity distri- bution companies were sold to local investors.12 Two hydroelectric plants were also given to AES under a 25-year concession.13 After protracted negotiations the bulk of Georgia's thermal genera- tion capacity was sold to AES in April 2000. Attempts to privatize the remaining generation and distribution assets outside Tbilisi have so far been unsuccessful. In April 2002 the government consolidated all remaining electricity distri- bution companies14 into the state-owned Georgian United Distribution Company (GUDC) to improve management and performance and make the remaining distribution assets more attractive to private investors. The government is hiring a private-sector management contractor for the GUDC; in the interim, day-to-day management oversight is being provided by consultants under USAID funding. The power transmission and dispatch companies (natural monopolies) were not offered for sale and remain in state hands (though an international management contractor has recently been engaged to operate power transmission and dispatch). In addition, the state retained ownership of the high-pressure gas transmission lines. Tbilgazi, the gas distribution company serving Tbilisi, was offered for privatization on a num- ber of occasions, but the only credible bidder has been Itera. The government regards Itera's ownership of Tbilgazi as an undesirable step towards the vertical re-integration of the gas supply sector, so Tbilgazi remains a municipally owned utility. The municipal government is proposing to engage a management contractor to operate the company. Energy Tariffs As part of the reform program electricity tariffs were increased to cost recovery levels.15 Prices of electricity have more than doubled since 1997 (in nominal terms).16 By contrast residential natural 11. Not including the Engurri Hydropower station which is located in the Abkhazia region. 12. Eight small companies (less than 5 percent of the market in total) in the Kakheti region have been sold. 13. Khrami I and Khrami II. 14. Outside the autonomous republic of Adjara, the autonomous region of South Ossetia, and the con- flict zone of Abkhazia. 15. These retail tariffs are the sum of the tariffs for electricity generation, imports, transmission, dispatch and distribution. Since 1998 the prices for transmission, generation, and distribution outside Tbilisi have been set by the National Energy Regulatory Commission (GNERC) on the basis of "cost plus return on investment", while distribution margins for Telasi have been set on the basis of escalation factors agreed in the purchase contract between AES and the Government. When the new tariffs were established in November 2002 (at 13.7 tetri/kWh), the government indicated that it proposed to use budget resources to pay the 10% increase, thereby leaving household tariffs in Tbilisi unchanged at 12.4 tetri/kWh. However, the Constitutional Court ordered a reduction in electricity tariffs on December 30, 2002. In compliance with this directive, GNERC issued a new order rolling back all prices by 10 percent (with the exception of elec- tricity from Enguri, where the price was reduced by 30 percent). AES Telasi has advised that it considers the roll-back to be a breach of their purchase agreement, and proposes to seek international arbitration. One possible response of the Government could be to use the budget funds already set aside in order to compen- sate Telasi for the difference between the "contractual tariff," and the reduced tariff set in accord with the directive of the Constitutional Court. 16. Prices in figure 4 are in nominal terms to reflect tariff increases, including those imposed by the reform. REVISITING REFORM IN THE ENERGY SECTOR 9 gas tariffs have remained relatively constant at GEL 0.27 per m3 in Tbilisi and GEL 0.30 per m3 in other cities. Households wishing to connect (or re-connect) to the gas network in Tbilisi must pay GEL 215 (equivalent to US$ 100 in 2002) to cover the cost of a meter17 or be billed GEL 6.50 per person per month.18 During focus group sessions, some participants noted the high up- front cost as a barrier to installing gas in their homes. Prices for non-network energy also increased substantially over the period--though not necessarily because of the reform program. For example, the large jump in kerosene price in 1999 may be related to rise in international crude oil prices, which rose from $10 to $22 per barrel between January and September. Relative fuel prices influence household energy consumption choices.19 The data suggest that clean network fuels--electricity and gas--have lower prices than non-network fuels--LPG and kerosene (figure 4). Even at full import prices, gas is substantially less expensive than all other fuels. While there may be additional costs associated with the technology required to use gas (metering and gas-fired appliances), the convenience and savings suggest that, given access, it is the household fuel of choice. Kerosene, an inferior fuel, is by far the most expensive and therefore least likely choice. An important omission in the comparison of energy prices is fuelwood--while commonly used, there is no reliable time-series data on price. Estimating wood prices is complicated by regional differences in availability (and thus in price), and by the fact that households can either collect wood or buy it whole or split. The HBS collects information only from households that have purchased wood, thus underestimating consumption. The STC survey found that, depending FIGURE 4: EFFECTIVE ENERGY PRICES, GEL PER MILLION BTU 80 70 Kerosene 60 LPG 50 (1997) 40 Electricity GEL 30 (Tbilisi) Electricity 20 (Regions) 10 Natural Gas 0 Jan-97 Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Source: Georgia State Department for Statistics. 17. Either up front or over time. 18. Gas tariffs at the end user level cover the cost of importing the gas from Russia (approximately US$60 per 1,000 m3), transmission charges, and the costs of local distribution. The transmission and distri- bution margins have been reviewed regularly by GNERC, and the companies are entitled to apply for a tariff increase based on demonstrated costs of service supply. 19. The prices shown are weighted national averages, based on data taken from the quarterly HBS. These prices are in cost per unit of effective energy output, rather than the prices that customers pay per unit of energy input. The adjustment was based on typical conversion efficiencies of the fuels, and the efficiency of different types of appliances (see annex A for details). This implicitly assumes that all households have the same technology. 10 WORLD BANK WORKING PAPER on the region, from 5 to 75 percent of households cut wood themselves. Even in this case there are important differences in access--and so in the time-related costs of collection. Recent survey research indicates that in the winter of 2002, wood prices were on the order of GEL 22 per m3. Assuming a typical conversion efficiency of 20 percent, the cost of wood energy would be GEL 15 per million BTU, less than all other fuels except natural gas. That makes wood the fuel of choice for cooking and heating for poor households not on the gas network. CHAPTER 3 IMPACT ON HOUSEHOLDS F or households, energy reform generally changes access (service quality) and prices. This leads to changes in consumption and expenditure patterns, including the risk that rising prices of network energy will cause consumers to increase their consumption of dirtier fuels, with negative externalities. This chapter explores the changes in household energy consumption and expenditure since 1996. Regional Differences in Energy Expenditures Despite rising prices, the household share of income spent on energy has remained constant at 8 percent.20 These national figures, however, mask differences between regions. The share of house- hold income spent on energy has increased most in Tbilisi (from 6.4 percent to 8.4 percent) and in other cities (from 6.9 percent to 8.7 percent), consistent with the privatization of Telasi and a shift to more expensive LPG in other major cities (figure 5). In rural areas the share of expenditures on energy remained almost constant until 2001. During 2002 expenditures in energy fell substantially due to a sharp decrease of kerosene con- sumption. Expenditures on electricity increased in rural areas, but the increase was not enough to offset the fall in kerosene expenditures. Expenditures on electricity are significantly higher in Tbilisi than in rural areas, consistent with the higher tariffs in Tbilisi and the regional trends in service quality. By the fourth quarter of 2001, 94 percent of households in Tbilisi received 24 hours of uninterrupted electricity, compared with 25 percent in other cities, and 7 percent in rural areas.21 20. This excludes 134 positive expenditures on coal. 21. Households in the HBS were asked to report the number of hours of electricity received during the week previous to the interview. Households were asked this question only during the first interview (house- holds were usually interviewed 4 times). The results shown above are for the quarter in which the initial interview took place. 11 12 WORLD BANK WORKING PAPER FIGURE 5: SHARE OF ENERGY IN TOTAL EXPENDITURES BY REGION 10% Rural areas 8% expenditures Urban areas 6% Tbilisi total in 4% energy 2% of Share 0% 1997 1998 1999 2000 2001 2002 Source: Georgia Household Budget Survey. Composition of Energy Expenditures Despite rising electricity prices the absolute value of expenditures on energy fell slightly (in real terms) over the period of analysis. One explanation is a reduction in the amount of energy used by households; another is substitution by less expensive fuels. Outside Tbilisi energy consumption has fallen since 1997, with some stabilization in mid-1999. The top quintile now consumes one-third as much energy (in effective BTUs) as in 1997 and the bottom quintile about half as much. Outside Tbilisi, fuelwood and kerosene remain significant in energy expenditures. Since kerosene is more expensive than electricity, it can be inferred that the consumption of kerosene is largely a response to inadequate electricity supply. Similarly the reduction in overall consumption can be attributed to budget constraints, and the lack of opportunity to substitute lower-cost fuels, such as electricity and natural gas, for kerosene and LPG. By implication, an improvement in elec- tricity and gas supply (keeping prices constant) is likely to result in welfare gains for households outside Tbilisi. In Tbilisi the top quintile's consumption initially dropped22 but eventually recovered to pre- reform levels, at about 200 million BTU per quarter--and the bottom quintile maintained the same consumption, at about 55 million BTU per quarter (figure 6). Relatively stable energy expenditure shares and consumption levels suggest that households in Tbilisi are replacing electricity with less expensive fuels. Breaking down total expenditures into its component reveals just this (figure 7). In Tbilisi households have increased the share of electricity in total energy from 45 to 51 percent from 1996 to 2002 (from 3 to 7 percent of income). The share of kerosene dropped. And shares of LPG and fuelwood (purchased) stayed constant. More significant, the share of gas increased from 2 to 20 percent, with the greatest increases in 1999. The focus group sessions examined the factors underlying these changing expenditure shares in greater detail, addressing the impact of access to gas on the energy mix of households. Most participants expressed a desire to have gas, preferring it to other fuels for both cooking and heat- ing--and to some extent for water heating. Participants noted that gas was cheaper than electrici- ty, cleaner and more comfortable than kerosene and wood. Almost all participants with no gas 22. Household fuel expenditures are converted into physical units (million BTU) by dividing expendi- tures by unit price per million BTU, and adjusting the physical units to reflect the conversion efficiencies of typical energy-consuming appliances (see annex A for additional details). REVISITING REFORM IN THE ENERGY SECTOR 13 FIGURE 6: TOTAL HOUSEHOLD ENERGY CONSUMPTION IN EFFECTIVE BTUS--TBILISI 350 300 250 Top Quintile BTU 200 150 Million 100 Bottom Quintile 50 0 II IV III II IV III II IV III II IV III II IV III II IV 1996-III 1997-I 1998-I 1999-I 2000-I 2001-I 2002-I Source: Georgia Household Budget Survey. connection said that they use kerosene or wood for heating and cooking. After getting access to gas, they give up these fuels. In fact, many said that they dislike both kerosene and wood so much that they use them only when no other option is available or affordable. Gas access gives them a desirable substitute. Installing a gas connection does not affect the level of electricity consumption, either because households were already controlling the use of electricity to reduce bills or because the areas where they live have supply restrictions. The latter applies to the areas with old and particularly non-working meters, where AES Telasi's losses are very high.23 FIGURE 7: ENERGY EXPENDITURE SHARES BY FUEL TYPE--TBILISI 60% 50% Electricity 40% 30% Kerosene 20% 10% Natural Gas 0% 1997 1998 1999 2000 2001 2002 Source: Georgian Household Budget Survey. 23. According to AES Telasi, in some areas estimates show that supply accounts for 60-70 lari per house- hold per month while payments are only 2-3 lari per household per month. 14 WORLD BANK WORKING PAPER Despite its desirable attributes, there are barriers to obtaining gas--mainly the costs of instal- lation and the meter and equipment. And to obtain a connection between the main pipeline and the building or residential area, it is necessary to have an agreement from all or most residents, not always easy to obtain. Some participants had been told that it was technically impossible to install the pipes in their area, and others have not yet been offered gas access. Changes in Electricity Consumption--Tbilisi The AES Telasi data allow detailed examination of household electricity consumption patterns over the last three years. Prices have increased, and customers are paying a larger share of their electricity bills. But mean household consumption has remained constant at around 125 kWh per month (figure 8), and median consumption constant at approximately 113 kWh.24 This reinforces the comments of the focus group participants: that gas is being used primarily as a substitute for wood, and that households limit their use of electricity owing to cost (and the obligation to pay) and to periodic supply limitations. The findings about mean consumption have two important policy implications. First, current consumption levels are low relative to what might be expected in urban areas in a country at Georgia's level of development. An average consumption of 125 kWh per month represents limit- ed use of electricity--for lighting and a modest number of appliances. It does not suggest exten- sive use of electricity either for heating or air conditioning.25 Second, demand in Tbilisi, where service has been quite reliable for the last few years, remains constant despite price increases, sug- gesting inelastic demand and large welfare losses from future price increases. FIGURE 8: HOUSEHOLD ELECTRICITY CONSUMPTION--TBILISI 200 Mean consumption 150 month 100 per Median consumption kWh 50 0 Jan-00 May-00 Sep-00 Jan-01 May-01 Sep-01 Jan-02 May-02 Sep-02 Source: AES Telasi. 24. The data set contains a large number of zeros during the first few months of 2000, so the median is close to zero. One explanation is that the billing system started in the middle of 1999, so the large number of zeros is part of the adjustment period during the creation of the data set. A second explanation is that there were few existing meters in the system during this period. Before new meters were widespread, an "average" or "estimated" amount of kWh was assigned to households as their consumption. These numbers were later verified by AES Telasi as new meters were introduced into the distribution system, sometimes resulting in very large bills for the households. 25. A refrigerator (manual defrost 5-15 years old) consumes about 95 kWh/month and 3 incandescent light bulbs another 30 kWh per month. REVISITING REFORM IN THE ENERGY SECTOR 15 Typically an electricity demand function is kinked, sloping steeply around the minimum required for basic needs and then rapidly leveling off as the quantity of electricity consumed moves from necessity to luxury. Identifying the location of the kink is important. At prices above this point, demand is inelastic and welfare losses associated with rising prices are large--while at prices below it, demand is more elastic and welfare losses are smaller. The distribution of annual household electricity consumption indicates that households have the highest probability of con- suming between 875 and 1,750 kWh a year (figure 9). Based on current consumption patterns of Tbilisi households, basic minimum needs are on the order of 1,500 kWh a year. Health and the Environment Among the key anticipated impacts of reform is that higher prices for clean network energy are likely to increase the use of dirty fuels (wood and kerosene) by the poor. Burning dirty fuel may be associated with significant negative externalities, including indoor and outdoor air pollution and deforestation. This may be an argument for keeping clean energy prices at a level that allows the poor to maintain access to them. The correlation between illness and household use of dirty fuels in badly ventilated homes is well established in the literature. The study examined household expenditure patterns on clean fuel (electricity, natural gas, and LPG) and dirty fuel (wood and kerosene). It found, as noted above, that households in Tbilisi have shifted to clean fuels (figure 10)--owing largely to increased supply of clean and inexpensive natural gas. This pattern holds for the bottom quintile and for the average household. Statistical analysis of the relationship between health outcomes (such as the incidence of acute respiratory infections) and fuel use did not reveal the same signifi- cant correlations that are picked up in larger time series data sets (see annex B). One reason may be the large number of confounding factors associated with observed health outcomes.26 Dirty fuel consumption in rural areas continues to pose a major public health risk. The STC survey indicates that 80 percent of rural energy consumption in the winter of 2001 was fuelwood. FIGURE 9: DISTRIBUTION OF ELECTRICITY CONSUMPTION--TBILISI, 2002 7% 6% 5% 4% 3% Frequency 2% 1% 0% 125 375 625 875 1125 1375 1625 1875 2125 2375 2625 2875 3125 3375 3625 3875 kWh per year Source: AES Telasi. 26. According to the STC survey, in 2002 over 53 percent of households had one or more members with a chronic disease, and 76 percent of households had one or more members with either an illness or disease in the previous three months. It is therefore possible that other factors for which there are no available data mask health differences related to fuel use. 16 WORLD BANK WORKING PAPER FIGURE 10: ENERGY EXPENDITURE SHARES ON CLEAN AND DIRTY FUELS--TBILISI 100% 80% shares Clean fuels (Electricity, Gas, LPG) 60% 40% expenditure Dirty fuels (Kerosene, Wood) 20% Energy 0% III III III III III III 1997-I 1998-I 1999-I 2000-I 2001-I 2002-I Source: Georgia Household Budget Survey. There may be welfare gains from increasing access to cleaner more efficient wood burning tech- nology. Figure 11 captures the variation in fuelwood prices and consumption level in different regions of the country. Improvements in technology could reduce the cost per effective BTU and increase the consumer surplus. Of course, it remains to be determined if households would adopt these technologies. Many other variables influence wood consumption, including forest cover, access to other fuels, proximity to forests, the availability of household labor to collect firewood (and, by extension, rural labor availability), and temperature. A more detailed household survey focusing on health outcomes, fuel use, the amount of ventilation, the type of equipment, the duration of fuelwood use, and the variables just mentioned might justify such an intervention. FIGURE 11: DEMAND FOR FUELWOOD IN REGIONS OF GEORGIA (WINTER 2002) 20 Adjara 18 Shida Kartli Kvemo Kartli 1 Samtskhe-Javakheti 2 16 Tbilisi Rustavi Kvemo Kartli 2 14 mmBTU Imereti Kakheti 12 Samtskhe-Javakheti 1 Mtskheta-Mtianet i 10 Samegrelo Racha-Lechkhumi Guria Effective 8 Svaneti per 6 GEL 4 2 0 0 20 40 60 80 100 120 Effective mmBTU from wood per year Source: Save the Children. REVISITING REFORM IN THE ENERGY SECTOR 17 Conclusions Improved service quality and increased supply of clean and inexpensive natural gas appear to have offset the potential negative impact of higher electricity prices in Tbilisi. The shift to natural gas also means that potential negative externalities from dirty fuel consumption may not have materi- alized. Regions outside Tbilisi, however, have not had the same opportunities for mitigation. These regions have seen the price of electricity increase and the availability of electricity supply decrease dramatically. A substantial number of households continue to burn wood, especially in rural areas. Improvements in wood burning technology that reduce the cost per BTU could pro- duce welfare gains. Households in Tbilisi are consuming about 125 kWh per month of electricity--close to basic minimum needs. Demand appears to be in the inelastic range, suggesting large welfare losses associated with future price increases. With a kinked demand curve, the welfare gains of providing large electricity subsidies to households (amounts greater than 150 kWh per month) are probably small. CHAPTER 4 IMPACTS ON UTILITIES T he impact of reforms on network energy suppliers has been mixed.27 The financial perform- ance of AES Telasi has unambiguously improved, but it remains unable to cover the costs of private capital--possibly owing to poor collection, high commercial losses,28 and higher than anticipated investment requirements. Corporate problems may have made this situation worse, as the sharp drop in share values has made it difficult for the parent to raise funds to cover invest- ments in subsidiaries, and perhaps to re-think some of its less-well performing investments. According to local management, Telasi's current position is not sustainable, and AES has on sever- al occasions indicated an intention to pull out of Georgia. Making matters worse are a roll-back in Telasi's approved tariff, ongoing disputes over nonpayment by budget enterprises, and Telasi's withholding of tax remittances. For the other companies in the electricity supply chain, collections have been poor--although GNERC has ensured that tariffs are at cost recovery levels.29 In many instances, GWEM has not collected enough revenue to cover the short-term operating costs--let alone the full costs--of some upstream service providers. So, despite relatively good performance by AES Telasi, the sustainability of the energy reform program is still in doubt owing to the poor financial condition of many of the enterprises. It is 27. While we are able to construct the performance of AES Telasi from the household consumption, billing, and payment data, we do not have equivalent data for the gas utility. Detailed statistics on gas sector arrears are limited. Sakgazi, the private company which owns and operates gas distribution companies in nine regional centers outside Tbilisi, was able to provide general numbers for collections in 2002. Collections overall averaged 81 percent, ranging from a low of 55 percent in Gori to over 93 percent in Bolnisi, Kaspi and Borjomi. Tbilgazi, the municipally owned gas distributor in Tbilisi advised that they were unable to pro- vide detailed collections data, since they were installing a new billings system. However, data collected by USAID consultants indicated that collections in the winter of 2002 averaged on the order of 25 percent. 28. Equal to about 40 percent of the energy purchased from the GWEM in 2002. 29. Including depreciation and operating costs plus a modest return on investment. 19 20 WORLD BANK WORKING PAPER hoped that the introduction of private management contractors to operate the distribution and transmission companies--as well as GWEM--will help to improve the flow of revenues, and hence the profitability and operating efficiency. However, it is too early to assess the likely impact of these initiatives. So the focus of this chapter is on how AES Telasi was able to improve revenues and what this means for GUDC and other enterprises in the supply chain and the gas sector. Performance of AES Telasi AES Telasi has dramatically improved revenues and cash flow since the beginning of 2000. Revenue from the residential sector increased 91 percent from 2000 to 2001 and another 41 per- cent from 2001 to 2002.30 While tariff increases account for some of the increase, better collec- tions from customers--as well as increases in the amount of targeted and non-targeted subsidies-- have also played a role. AES Telasi has been particularly successful at reducing household arrears. Over time there has been steady improvement in collection rates, rising from 44 percent in 2000 to 86 percent in 2002. At times collection rates have even exceeded 100 percent of current billings, as households settled arrears and transfer payments for subsidies were received from USAID or the government. The key instruments used by Telasi to achieve improved collection rates from customers include better service quality and metering. The data suggest that metering and subsidies had a much larger impact on collection rates and revenues than service quality and retail prices (table 1).31 Two tools were used to identify to role of the different instruments in Telasi's performance. First, a multivariate analysis estimated receipts as a function of service quality (ratio of requested and received energy), price, enforcement (percentage of households that have been re-metered), and subsidies. The analysis also controlled for monthly temperature and the temporary loss of ther- mal power plants in the winter of 2001. Model details are summarized in annex C. The multivariate results indicate that re-metering and price are equally important determinants of receipts, followed by quality and subsidies. The analysis also shows that collections increase at a decreasing rate with re-metering, suggesting that re-metering may yield higher revenues than prices in the initial stages of reform. As mentioned earlier, the cost of meters is not taken into account in this analysis. The second tool was the focus group sessions, which solicited the views of Telasi customers on a wide range of issues related to improved payment levels, including re-metering, enforcement, and service reliability. Prices It is difficult to untangle the role of prices from that of enforcement and service quality in improv- ing AES Telasi's revenues. Higher prices would be expected to increase revenues. But price increases can lead to reduced consumption, and possibly to increased nonpayments. The simple tabulations in table 1 indicate that both consumption and collection rates improved in conjunc- tion with recent tariff increases. Receipts increased 44 percent from 2001 to 2002 as prices rose 24 percent. In the previous period receipts increased 91 percent as prices rose 8 percent. 30. These figures are for a sample of 1,349 households included in the Georgia Household Budget Survey. In total, AES Telasi has approximately 300,000 customers. Households participating in the HBS were randomly selected, and may be presumed representative of the total population of households in Tbilisi. 31. The cost of meters is not taken into account in the analysis. REVISITING REFORM IN THE ENERGY SECTOR 21 TABLE 1: AGGREGATE IMPACT OF REFORM ON COLLECTION RATES--TBILISI Change Change 2000 2001 2002 '01 vs '00'02 vs '01 Telasi received power--million kWh 2.79 2.38 1.20 -15% -6% Telasi requested power--million kWh 3.23 2.76 1.29 -14% -20% Ratio of received to requested power 86% 86% 93% 0 pp 7 pp Average price (GEL/kWh) 0.093 0.100 0.124 8% 24% Portion of households re-metered 38% 69% 76% 32 pp 7 pp Consumption--million kWh 2.35 2.31 2.49 -2% 24% Billings--thousand GEL 217 232 309 7% 33% Total receipts--thousand GEL 96 186 266 93% 44% Subsidies--thousand GEL 35 44 55 25% 26% WHAP 29 37 47 28% 27% Government privileges 6 7 8 11% 21% Payments by customers--thousand GEL 61 142 211 132% 49% Collection rate from households 44% 80% 86% 36 pp 6 pp Notes: Table includes only Tbilisi households in the sample. Requested and received power in 2002 covers from January to June only. pp = percentage points Source: Data from AES Telasi. Subsidies Subsidies are important for AES Telasi. The Winter Heat Assistance Program32 (WHAP) account- ed for 29 percent of receipts in 2000 and about 18 percent in 2001 and 2002. Government privi- leges account for anywhere between 3 and 6 percent of AES Telasi's receipts in a given year. The revenue from subsidies grew in absolute terms--largely owing to the increasing WHAP benefit (in both number of kWh provided and also the associated tariff). In terms of revenue shares, however, the importance of subsidies decreased because of the large increase in collections from households. Service Quality A reasonable proxy for service quality is the hours of service that consumers receive. Because the data needed to relate aggregate hours of supply within the AES Telasi service area to hours of service for individual customers were not available, it was not possible to study in detail how changes in hours of service affected individual payment rates and arrears. It was possible to com- pare collection rates with the ratio of received to requested power, but there was no substantial correlation, possibly because Tbilisi receives close to 24 hours of service a day. 32. The Winter Heat Assistance Program (WHAP) is administered and largely financed by USAID. This program finances the supply of electricity to low-income households for winter heating during the January- April period. The amount each household receives has varied each year depending on the funding available. It was 850 kWh in 2000 and 1000 kWh in both 2001 and 2002. The planned amount for 2003 is 480 kWh. The WHAP is focused on electricity customers in Tbilisi (where tariff increases have been highest). But part of the funding is allocated to other urban centers. The program is entering its fifth year. 22 WORLD BANK WORKING PAPER Reliability of supply did not seem to be a major direct factor affecting the payment patterns of focus group participants, but most participants noted that service quality had improved signifi- cantly since Telasi's privatization. Some participants also noted that they were anxious to get new meters because "supply is better when you have them." Some participants also expressed dissatis- faction with Telasi's failure to adhere to its original promise that if customers paid their bills, they would have 24 hours of improved electricity service. This suggests that service quality may have affected the payment patterns of some households. One of AES Telasi's most vexing problems has been the difficulty of obtaining enough power to meet winter demand in Tbilisi. This problem was partly solved by refurbishing the distribution network, by entering into private contracts with suppliers, and taking owner-manager positions with generators. In essence AES Telasi bypassed the electricity market created by the reform pro- gram--it re-integrated vertically to counter corruption and inefficiency in the segments of the supply chain not under its control. Similarly, the Russian gas exporter, Itera, holds either full or controlling ownership of Sakgazi. While this de facto vertical re-integration of the electricity and gas supply sectors contravenes at least one of the initial objectives of sector unbundling--promot- ing competition in supply--it may be a pragmatic and necessary step when privatization is incom- plete or very small. Enforcement In the statistical analysis, enforcement explains much of the improvement in collections. With re- metering33 as a proxy for enforcement, collection rates are systematically higher for re-metered households. There is no statistically significant difference in consumption between re-metered households and those that have old meters (figure 12). Re-metered households pay a systematical- ly higher percentage at all consumption levels.34 Similarly, arrears are significantly lower for re- metered households. FIGURE 12: COLLECTION RATES BY RE-METERED STATUS--TBILISI 175% 150% Re-metered Households 125% rate 100% 75% Collection 50% Households 25% not Re-metered 0% Jan-00 May-00 Sep-00 Jan-01 May-01 Sep-01 Jan-02 May-02 Sep-02 Source: AES Telasi. 33. Re-metering refers to both replacing old meters for newer ones, and installing it outside of the dwelling; households used to have meters inside the dwelling but they've been replaced for new ones. 34. Re-metered households pay on average twice as much as those not yet remetered. REVISITING REFORM IN THE ENERGY SECTOR 23 A shortcoming of the statistical analysis is that it does not provide guidance on whether metering facilitates cut-offs for nonpayment (enforcement) or adds credibility to the invoice (a proxy for service quality). So, the interaction between metering and payments was a key issue addressed in the focus groups. The responses of participants indicated that the role of metering is complex. Participants feared supply cut-offs, trusted that the amount of their bills was accurate (though some participants expressed doubts about the accuracy of the new meters, which appeared to "go faster"), and controlled consumption and hence the amount owed for electricity. Some participants also noted a reduction in corruption as a result of the new meters as an advan- tage, though others appeared to see this as a negative. The fear of cut-off was particularly strong--even though Telasi has advised that they probably cut off only 10 percent of nonpaying households in each month. This suggests that the threat of disconnection (particularly if it is likely to occur at an inconvenient time) may be almost as effec- tive in reducing nonpayment as an actual cut-off. In addition, participants who paid their bills expressed resentment that Telasi did not do a better job tracking down and removing illegal connections. Structure of Arrears Improving collections might have a disproportionate impact on low-income households. But a comparison of changes in collection rates by income class indicates that they have increased uni- formly across the top and bottom quintiles (figure 13). This challenges the conventional wisdom that nonpayment is more closely related to affordability than free-riding. If affordability were more important, collections would be expected to be lower for the bottom quintile. Conclusions The sustainability of the reform program is still in doubt, despite very good performance by the main private operator. Over the last three years AES Telasi has increased receipts 135 percent. The data suggest that re-metering is as important a determinant of utility receipts as prices, followed FIGURE 13: COLLECTION RATES BY QUINTILE--TBILISI 175% 150% Top Quintile 125% 100% 75% 50% Bottom 25% Quintile 0% Jan-00 Jul-00 Mar-00 May-00 Sep-00 Nov-00 Jan-01 Jul-01 Mar-01 May-01 Sep-01 Nov-01 Jan-02 Jul-02 Mar-02 May-02 Sep-02 Nov-02 Source: AES Telasi. 24 WORLD BANK WORKING PAPER by service quality and subsidies. The multivariate analysis suggests that a high priority should be placed on re-metering in conjunction with tariff increases, particularly at the early stages of reform, to generate the maximum amount of revenue. The data also suggest that an aggressive approach to reducing nonpayment does not necessari- ly have a disproportionate adverse impact on low income households--particularly if suitable sub- sidy and transfer mechanisms address cases of severe need. CHAPTER 5 IMPACTS ON GOVERNMENT T he fiscal impact is one of the central arguments to support energy reform. The short-term impact is through privatization receipts. The long-term impact on the budget is a more inter- esting issue, but one that has been little studied despite its importance. In Georgia energy sector reform has so far increased government expenditures on the sector. Part of this increase arises from improved transparency and the monetizing of formerly implicit subsidies and arrears. But expenditures for energy have also increased as a result of tariff increases, which mean the government must pay more for households receiving subsidies, as well as for ener- gy consumed by organizations supported by the budget. Government expenditures for gas have gone up as Tbilgazi brings more residential customers on line without improving cost- recovery. A comparison is made between the existing categorical subsidies and the Winter Heat Assistance Program (WHAP) and a new proposal for targeting the existing subsidy based on household elec- tricity consumption. State Support for the Energy Sector While GDP began to recover after 1994, the increases were accompanied by a more-than-propor- tional increase in budget expenditures. From 1995 to 2000 government expenditures (including net lending) rose from 12 percent of GDP to 19. Notwithstanding revenues received from privati- zation, ongoing payments to the utilities were a significant and growing portion both state and municipal budgets. Identifying the precise amounts that the government spent on the energy sec- tor between 1991 and 1999 is impossible because accounts are not fully segregated. But over this period the state and the energy sector jointly accumulated GEL 287 million in external debt for the supply of electricity and fuel, and GEL 354 million in internal debt to the budget, suppliers, and commercial banks, as well as internal arrears among sector enterprises. Borrowings from 25 26 WORLD BANK WORKING PAPER donor and bilateral agencies (EBRD, IDA, KfW, OECD countries) over the period totaled an additional GEL 202 million.35 More recent disaggregated data indicate that central government expenditures continue to rise, totaling more than GEL 210 million from 2001 to 2003 (table 2). In addition to direct budget support, the government provides indirect support through the deferral of VAT on the accumulated arrears of the sector, and the waiver of VAT on technical and non-technical losses. The government also provides guarantees and co-financing for energy sector investment projects financed by bilateral and multilateral financing agencies. The accumulation of external arrears has slowed, but interenterprise debts within the sector have accelerated, creating a potential govern- ment liability for state-owned enterprises and reducing tax revenues. At the end of 2002, total debts of the GWEM to suppliers (generation, transmission) and to the budget were GEL 444 million. Direct budget expenditures for energy have increased from 43 million GEL in 2001 to an expected 98 million GEL in 2003--equivalent to 7.3 percent of total government expenditures (table 2). A substantial portion of the increase (22 million GEL) is the growth in payments by the government to the energy sector as partial compensation for electricity consumed, but not paid for, in Abkhazia and Tskhinvali Region. Payment of this amount is an important step forward in terms of monetizing a formerly hidden subsidy. A second significant component of the expenditure increase consisted of subsidies to Tbilgazi to allow it to settle debts to suppliers and ensure the flow of gas to Tbilisi. The remainder of the increase in expenditures from 2001 to 2003 is due primarily to higher electricity costs. This increase affects government expenditures in three ways. First, as electricity costs go up, subsidies to refugees and other households cost more. Second, the government has budgeted to take on the financial burden of recent tariff increases owed to Telasi. Third, the government must pay more for the electricity it consumes. Municipal Support for the Energy Sector Although little information is available on local government expenditures for energy outside Tbilisi, municipally provided subsidies in the city have increased. Subsidies to Tbilgaz have climbed sharply since 1999, reaching 10 million GEL in 2002 (table 3). The central government TABLE 2: STATE BUDGET PAYMENTS TO THE ENERGY SECTOR 2001-2003 (THOUSAND GEL) 2003 Name 2001 2002 (Plan) Direct subsidy to the Ministry of Fuel and Energy 3,000 13,000 36,500 Reimbursement of the fee for electricity consumed by the refugees 6,555 13,646 14,016 Reimbursement of the fee for electricity consumed by the budget organizations (Central, Local) 21,924 27,346 29,348 Sums allocated for energy sector through special decrees 6,000 10,160 4,500 Compensation for the various categories of population 2,800 3,000 11,500 Total direct support 42,280 69,154 97,867 Total budget expenditures 906,314 1,031,259 1,343,700 Energy sector support/total budget (percent) 4.7 6.7 7.3 Foreign credits and cofinancing 17,279 34,325 46,500 Source: Ministry of Finance. 35. This was largely at concessionary repayment terms. REVISITING REFORM IN THE ENERGY SECTOR 27 TABLE 3: STATE BUDGET ENERGY SUBSIDIES--TBILISI (THOUSAND GEL) 1998 1999 2000 2001 2002 Gas subsidies 50 15 340 842 10,113 Electricity subsidies 495 0 0 0 0 Total 545 15 340 842 10,113 Total expenditures 107,916 123,989 127,190 154,760 182,686 Subsidies/total 0.5% 0.0% 0.3% 0.5% 5.5% Source: Tbilisi Municipality. provided an additional 10 million GEL to Tbilgaz. Municipal subsidies for electricity were stopped in 1999 with the privatization of Telasi. In 1996 only 10,000 households in Tbilisi were connected to gas, but this number increased significantly after 1998 and has now reached 170,000 households. High technical and commercial losses as well as the large number of households eligible for subsidized gas have meant that subsi- dies grew as new customers were added.36 The need for subsidies will continue to grow until the company's performance is improved. Electricity Subsidy Effectiveness Energy subsidies to Georgian households are available through a range of programs. One govern- ment program provides all veterans and pensioners between 35 and 70 kWh per household a month (recently increased to 240 kWh a month in the winter and 120 kWh a month in the sum- mer). Refugees and internally displaced persons (including those not living in collective centers) also receive substantial quantities of free electricity. Other government programs provide house- holds 850 m3 of natural gas a year.37 In addition to the government-funded subsidies, a major donor-financed subsidy to electricity customers­WHAP­has been in place for the past five years.38 Data from AES Telasi were matched to corresponding household data from the HBS to examine the targeting of the two electricity subsidies in Tbilsi (government subsidy and WHAP). The percentage of households receiving the government subsidy (paid to veterans and pensioners, and not specifically poverty targeted) is evenly divided across all quintiles (table 4). The WHAP subsidy, which is poverty targeted, accrues more to households in the lower quintiles. Even so, a significant share of the WHAP accrued to households in the high expenditure quintiles in 2000 and 2001. To improve the targeting, USAID is reviewing the eligibility of households receiving assistance under the program. Given earlier findings about basic minimum needs and average household electricity consump- tion, it is worth examining how much of consumption is covered by the subsidies. Recipients of the government subsidy get 27 to 32 percent of their annual electricity for free (table 5), and those of the WHAP subsidy, 54 to 64 percent. More detailed analysis suggests that WHAP recipi- ents do not necessarily use the free electricity for heating--in many cases, they use the subsidy to smooth their consumption through the entire year. This may explain in part how households have 36. Households are eligible for gas subsidies according to categorical privileges. Gas subsidies come both from central government and from the municipality (in Tbilisi). 37. This program is part of the "President's fund" which covers veterans. 38. As noted earlier, the Winter Heat Assistance Program finances the supply of electricity to low-income households in Tbilisi for winter heating during the January-April period. The amount each household receives has varied each year depending on the funding available. It was 850 kWh in 2000 and 1000 kWh in both 2001 and 2002. The planned amount for 2003 is 480 kWh. The future of this program is unclear now that AES has withdrawn from Georgia. 28 WORLD BANK WORKING PAPER TABLE 4: ELECTRICITY SUBSIDY INCIDENCE--TBILISI Quintile Year Bottom Mid-Low Middle Mid-Hi Top Percentage of households receiving government subsidy 2000 12% 12% 15% 13% 13% 2001 10% 16% 18% 11% 10% Percentage of households receiving WHAP subsidy 2000 25% 16% 18% 17% 10% 2001 27% 23% 21% 19% 14% Source: Georgia Household Budget Survey and AES Telasi. managed to maintain (and sometimes even increase) electricity consumption despite substantial tariff increases. Alternatively the targeting could be based on a rolling average of household consumption (say, in the three previous months). But because there is surprisingly little differentiation in con- sumption between the bottom and top quintile during summer months, a simple simulation per- forms better if based on annual consumption (figure 14). The proposed subsidy would be given to households consuming between 875 and 1750 kWh a year--or between 73 and 145 kWh a month. The lower bound is set in order to exclude residences, such as summer houses, that are not occupied on a regular basis. The proposed subsidy would reach a higher percentage of low-income households than either of the existing subsidies (table 6). It would also reach a higher percentage of the other quintiles as well. The absolute subsidy to each household would be substantially lower than in either of the existing programs. The total cost would fall between the WHAP and the government program. The new program would thus be more cost-effective (in GEL per household) than either of the existing programs. The simulation illustrates an alternative subsidy design, but there are several important caveats. First, the cost of the proposed subsidy would increase as the old subsidy is phased out, reducing of some of the fiscal benefits. This is because more households are likely to consume in the 75 to 125 kWh range. At the same time, poverty targeting may well improve as the old sub- sidy is phased out. With the loss of the existing subsidies, consumption will be based more directly on actual household income. Second, there are several well-organized stakeholders encouraging the government to keep the subsidies in place, including veterans (who do not wish to lose their benefits) and Telasi (which presumably enjoys the simplicity and predictability of payments associ- ated with the current system). Third, these results are based on data from Tbilisi and caution must be taken in generalizing them to the rest of the population. TABLE 5: SUBSIDY COVERAGE--TBILISI Government subsidy USAID subsidy (WHAP) Mean annual (kWh) % kWh free Mean annual (kWh) % kWh free 2000 1851 28% 1440 54% 2001 1659 32% 1461 64% 2002 1948 27% 1691 56% Source: Georgia Household Budget Survey and AES Telasi. REVISITING REFORM IN THE ENERGY SECTOR 29 FIGURE 14: FREQUENCY OF HOUSEHOLD ELECTRICITY CONSUMPTION (KWH PER YEAR)--TBILISI 8% Bottom Quintile 6% 4% All other Quintiles Frequency 2% 0% 125 375 625 875 1125 1375 1625 1875 2125 2375 2625 2875 3125 3375 3625 3875 kWh per year Source: Georgia Household Budget Survey and AES Telasi. TABLE 6: SIMULATION OF SUBSIDY COST-EFFECTIVENESS--TBILISI Quintile Bottom Mid-Low Middle Mid-Hi Top Households receiving (percent): Government subsidy 10 16 18 11 10 WHAP subsidy 27 23 21 19 14 Proposed subsidy(a) 44 38 40 42 39 Proposed subsidy--no gas users(b) 40 35 43 34 35 Average subsidy per household (kWh/yr) Government subsidy 610 561 548 646 535 WHAP subsidy 1000 1000 1000 1000 1000 Proposed subsidy(a) 407 411 497 476 324 Proposed subsidy--no gas users(b) 398 384 479 382 287 Cost effectiveness (GEL/household) Government subsidy 76 70 68 80 66 WHAP subsidy 124 124 124 124 124 Proposed subsidy(a) 50 51 62 59 40 Proposed subsidy--no gas users(b) 49 48 59 47 36 Notes: (a) The proposed subsidy is for households that consume between 875 and 1750 kWh a year. These households are given a monthly subsidy equal to the difference between 125 kWh and their monthly con- sumption. The assumed tariff is 0.124 GEL/kWh. (b) The second proposed subsidy is identical to that described in (a), except that it is available only for households that do not have access to natural gas. Source: World Bank. 30 WORLD BANK WORKING PAPER Nonetheless, one possibility would be to pilot the new subsidy program as part of the new management changes for GUDC. The HBS could be linked directly to the utilities' billing and payment database to monitor the poverty targeting of the subsidy. Over time--as data on con- sumption patterns, income, and payment are collected and analyzed--the targeting system could be refined to reduce the overall cost. Conclusions Government expenditures on the energy sector, as recorded in the consolidated budget, have increased since the implementation of sector reforms. Many of these expenditures (such as pay- ments to the electricity market for electricity consumed by Abkhazia and expenditures of budget enterprises for energy supply) simply recognize costs that were incurred but not paid before the reform. But government expenditures on subsidies are also high and growing. The rising cost of electricity subsidies can be attributed both to rising tariffs and to government decisions to increase the support (in both the number of kWh provided under specific programs, and the coverage of contractually mandated increases in Telasi's tariff). In addition, the cost of subsidies for gas supply (as provided by both the state and municipal budgets) is rising as additional households eligible for support of energy expenditures are connected to the gas network. Because subsidies can be a tool of poverty alleviation (and hence increasing equity), the merits of the current system are dubious. A significant part of the subsidies go to households in the high- er expenditure quintiles. In addition, a large share of the subsidies--at least under the expanded program of privileges--is compensation for electricity consumption beyond levels that might be considered "basic." This suggests that the government is, in many instances, financing consump- tion in excess of what typical households would be willing to consume if they were obliged to pay from their own household budgets. One of the most contentious debates in the power sector is between tariff based subsidies and direct income transfers. Proponents of direct income transfers argue that lifelines are not targeted and thus encourage inefficient energy use. Opponents claim that transfers through the general social assistance system, while theoretically attractive, do not take into account externalities and fail to reach a large share of the poor because of inadequate targeting. In Georgia, the welfare gain to households associated with the mis-direction of subsidies is small, but the burden on the government budget precludes the payment of other types of benefits that could have a substantial impact on welfare--particularly for low-income households. Re- orientation of the subsidy program towards maximizing the compensation for welfare losses would benefit both the consumers and the government budget. One solution would be to base targeting on electricity consumption, and provide enough compensation to ensure that the household receives a basic level of electricity needs. Such a subsidy program would provide a relatively simple mechanism for targeting--and would be more cost-effective in welfare gains per unit of subsidy paid. ANNEX A CONVERTING ENERGY PRICES INTO COST PER EFFECTIVE BTU D ata are not available on energy consumption at the household level. This information is inferred from the amount households reported paying for each fuel. The quantity of energy consumed by each household is estimated by dividing the amount reported paid by the fuel's cost per effective BTU. This method has several shortcomings, because households sometimes con- sume energy sources which they do not pay for. For instance, households may collect their own wood, so their energy expenses for wood would be zero. Similarly, households may not pay for electricity they receive. In this sense the quantity of energy estimated will be lower than the total consumed by the household. The method to estimate each fuel's cost per effective million BTU (mmBTU) is shown in table A.1. This table contains energy prices for December 2002 in Tbilisi, but the same method was applied to all regions across time. The fuel's original cost (column 3) is divided by its energy content (column 4) and divided by 106 obtain the cost in million BTUs (mmBTU). Wood is the cheapest fuel, followed by natural gas, while electricity is the most expensive. Column 6 represents the household technology. The cost per effective mmBTU takes technology into account. The lat- ter cost implies, for instance, that it is more expensive to heat the same space using wood than gas because of the efficiency of each appliance. Column 7 shows that gas is the cheapest fuel in effec- tive mmBTU, followed by wood and electricity, while kerosene is the most expensive. 31 32 WORLD BANK WORKING PAPER TABLE A.1: CALCULATION OF COST PER EFFECTIVE BTU Household Energy price in content Cost per Tbilisi, [BTU per Cost per Efficiency effective Dollars per December original mmBTU (household mmBTU effective Fuel Original 2002(a) unit](b) (GEL) use)(c) (GEL) mmBTU(d) [1] [2] [3] [4] [5]=10-6[3]/[4] [6] [7]=[5]*[6] [8] Natural Gas m3 0.270 3,412 7.65 70% 10.93 $ 5.08 Electricity kWh 0.137 35,300 40.15 90% 44.61 $20.75 Kerosene liter 0.790 32,934 24.04 40% 60.09 $27.95 LPG kg 1.400 42,854 32.67 70% 46.67 $21.71 Fuel wood m3 22.563 7,165,200 3.15 20% 15.74 $ 7.32 Notes: a. Energy prices (except wood) from State Department of Statistics. Price of wood from USAID/Save the Children. b. Mission estimates. c. Mission estimates. d. Exchange rate was 2.15 in December 2002. Source: Author's calculations. ANNEX B ENVIRONMENTAL OUTCOMES S tatistical analysis of the relationship between health outcomes (such as the incidence of acute respiratory infections) and fuel use did not reveal significant correlations that are often picked up in larger time series data sets. Both the Save the Children and Georgia Household Budget Surveys include questions on health status of household members. Correlation tests and regression analysis was used to examine the possible correlation between respiratory infections and fuel use and between gastrointestinal problems and access to piped water. No evidence was found of a positive and significant correla- tion between use of dirty fuels and respiratory infections. Next is a list of the correlations tested using both correlation tests and regression analysis. Acute Respiratory Infections Heating with wood and someone in household suffers from acute respiratory infections (ARIs) Heating with dirty fuels and someone in household suffers from ARI Heating with wood and children suffering ARI Heating with dirty fuels and children suffering ARI Heating with wood and adults over 60 suffering ARI Heating with dirty fuels and adults over 60 suffering ARI Cooking with dirty fuels and someone in household suffers from ARI Cooking with wood and someone in household suffers from ARI Cooking with dirty fuels and females suffering ARI Cooking with wood and and females suffering ARI Cooking with dirty fuels and females under 18 suffering ARI Cooking with wood and and females under 18 suffering ARI 33 34 WORLD BANK WORKING PAPER Cooking with dirty fuels and adults over 60 suffering ARI Cooking with wood and and adults over 60 suffering ARI This exercise was repeated when households were either cooking or heating with dirty fuels (this is, using a dummy variable equal to 1 if household heats or cooks with dirty fuels). Furthermore, these correlations were also analyzed by region and by building type (for example, apartment buildings in Tbilisi that heat and cook with dirty fuels) ANNEX C ANALYSIS OF TELASI'S REVENUES T his annex deals with the question of which has had a greater effect on Telasi's revenues: an increase in electricity tariffs or an increase in enforcement. Table 1 in the report suggests that there is a difference in the effect that these two forces have had on Telasi's revenues. Between 2000 and 2001 the price of electricity increased by 8 percent, the portion of households re- metered39 increased by 32 percentage points and revenues increased by 91 percent. But between 2001 and 2002 the average price of electricity increased by 24 percent, rate of re-metering increased by 7 percentage points, and total revenues increased only by 44 percent. These facts may suggest that enforcement (represented by re-metering) had a larger impact on revenues than the tariff increase, but many other factors affect this result. Collection rates were very low during 2000, so a small increase in enforcement may have produced a larger increase in revenues than if the collection rates were higher. And according to Telasi, the quality of service has increased over time. Unfortunately, we don't have a variable to measure this--we have infor- mation only on the requested and received power. The remainder of this annex tests whether enforcement of tariff increases have been more beneficial to Telasi. We propose a simple multivariate econometric model of revenues to address the question above. Revenues depend on the amount on kWh sold, the price at which they were sold, and the collection rate. The electricity tariff is determined by the GNERC and is exogenous to Telasi. The total kWh sold depend both on the total demand (from households, commercial sector and indus- try) and the total amount supplied to Telasi by the system. This amount is very seasonal and 39. Re-metered households are those in which Telasi has installed a new meter outside of the dwelling. In this analysis we assume that enforcement can be proxied by the re-metering status of the household. There are several problems with this approach--households may tamper with the new meter making their bills lower; Telasi may not be able to disconnect non-paying households or force households to pay their electricity consumption. Nevertheless, re-metering is a necessary condition to identify households not paying for their utilities. Therefore the threat of disconnection (enforcement) is very linked to re-metering. 35 36 WORLD BANK WORKING PAPER increases during the winter months, therefore temperature is included in the model as an explana- tory variable. Also, a closer look at the data suggests that something out of the ordinary happened during the first quarter of 2001. During these days there was a shortage of supply because of problems in the generation system. Revenues are also determined by the collection rate. It can be argued that collection rates depend both on enforcement and quality of service. As enforcement, or the threat of disconnec- tion, increases, households are more likely to pay for the electricity they consume. Households may not pay their entire bill, but they may pay a portion that guarantees they are not disconnect- ed. As of this date, households in Tbilisi are required to pay the full bill within fifteen days of receipt in order not to be disconnected. There is still a large portion of households not paying their bill in full, but Telasi has capacity to disconnect only a small portion of them. In some cases, Telasi workers disconnect houses late at night for fear of violence. Finally, collection rates may be influenced by quality of service. Households some times may not feel entitled to pay because the service is not as good as they expect. This argument is in part supported by the data (table C.1) there exists a positive and significant correlation between hours of service and amount paid for electricity both in urban areas outside Tbilisi and in rural areas. Also, the average amount paid is statistically different between rural and urban areas. We estimate a system of equations where revenues and collection rates are endogenous. In its reduced form, the revenue function will be a function of prices, the collection rate, the level of subsidies, and temperature. As mentioned before, the collection rate is a function of enforcement, and quality of service. Controlling for the winter of 2001, the final model to be estimated is shown in table C.2: The model was estimated using the instrumental variable regression procedure in STATA (table C.3). The model was estimated using time series data. Monthly data is available for three years for 36 observations. The fit of the model is good, as shown by the R-squared and the F-test on the top of the table. All coefficients except for that of the average temperature are significant. A Hausman test was conducted to test whether there is sufficient correlation between the distur- bances 1 and 2 to warrant estimation by instrumental variables. The Hausman test indicated that OLS is an inconsistent estimator for equation 1.40 TABLE C.1: AVERAGE HOURS OF ELECTRICITY AND AMOUNT PAID BY REGION Tbilisi Other urban Rural Hours of Amount Hours of Amount Hours of Amount Quarter service paid service paid service paid 2000-I 13.9 5.3 9.6 4.5 7.1 2.3 2000-II 13.9 4.6 11.1 4.8 7.4 3.1 2000-III 23.1 5.9 10.9 4.6 9.4 3.0 2000-IV 20.7 6.6 9.9 5.3 9.6 3.6 2001-I 11.4 7.9 6.9 4.4 4.3 2.0 2001-II 12.3 8.3 7.5 5.6 6.9 3.4 2001-III 22.6 9.4 9.6 4.8 9.1 4.0 2001-IV 23.5 14.2 12.7 5.8 8.6 4.0 Source: Household Budget Survey 40. The chi-squared for the Hausman test equals 6.73 and the null hypothesis that OLS on equation 1 is not an inconsistent estimator can be rejected at the 1 percent level. Test: Ho: difference in coefficients not systematic chi2( 1) = (b-B) ' [(V_b-V_B) ^ (-1)](b-B) = 6.73 Prob>chi2 = 0.0095 REVISITING REFORM IN THE ENERGY SECTOR 37 TABLE C.2: REGRESSION MODEL ln R = 0 + 1 ln Coll + 2 ln P + 3 Temp + 4 Winter01 + 5 ln Subsidy + 1 [1] ln Coll = 0 + 0 enf + 1 enf2 + 2 quality + 2 [2] where ln R ­ log of revenues ln Coll ­ log of collection rate ln P ­ log of price Temp ­ average monthly temperature Winter01 ­ dummy equal to 1 representing the winter months of 2001 ln Subsidy ­ log of subsidy enf ­ enforcement enf2 ­ enforcement squared quality ­ quality i ­ represent error terms It is expected that revenue will increase with collection rates, price increases, and subsidies received by the company. The results support this belief. Increasing collection rates by 1 percent will improve revenue by 0.7168 percent and increasing price by 1 percent will increase revenues by 0.6565 percent. The difference between these two coefficients is not statistically significant. Similarly, increasing subsidies by 1 percent increases revenues by 0.2555 percent. Finally, revenues increase when temperature decreases since a greater consumption of electricity can be expected for heating purposes and the winter of 2001 brought about lower consumption due to a lower avail- ability of power for sale. TABLE C.3: REGRESSION RESULTS Regression on Revenue Instrumental variables (2SLS) regression Source SS df MS Number of obs = 36 F( 5, 30) = 82.64 Model 7.01960499 5 1.403921 Prob > F = 0.0000 Residual .517717153 30 .017257238 R-squared = 0.9313 Adj R-squared = 0.9199 Total 7.53732214 35 .215352061 Root MSE = .13137 ln Revenue Coef. Std. Err. t P>|t| [95% Conf. Interval] ln Coll .7167994 .0784598 9.14 0.000 .5565631 .8770357 ln P .656545 .3602119 1.82 0.078 -.079106 1.392196 Temp_avg -.0030601 .0031612 -0.97 0.341 -.0095162 .0033959 Winter01 -.1818197 .0964468 -1.89 0.069 -.3787904 .015151 ln Subsidy .2666331 .0260078 10.25 0.000 .2135182 .3197481 Constant 9.417039 1.025128 9.19 0.000 7.323448 11.51063 continued on next page 38 WORLD BANK WORKING PAPER TABLE C.3 CONTINUED Regression on Collection Rate (Instrumented) . reg lncoll enf enf2 quality Source SS df MS Number of obs = 36 F( 3, 32) = 93.61 Model 7.84100824 3 2.61366941 Prob > F = 0.0000 Residual .89351395 32 .027922311 R-squared = 0.8977 Adj R-squared = 0.8881 Total 8.73452219 35 .249557777 Root MSE = .1671 lncoll Coef. Std. Err. t P>|t| [95% Conf. Interval] enf 2.571911 .5743944 4.48 0.000 1.401907 3.741914 enf2 -1.213233 .7249909 -1.67 0.104 -2.689991 .2635254 quality 1547864 .2499061 0.62 0.540 -.3542556 .6638284 _cons -1.766414 .2681718 -6.59 0.000 -2.312662 -1.220165 The effect of enforcement on revenues can be recovered by using equations 1 and 2. The sec- ond set of results in the bottom of table C.3 represents the regression of the log of collection rate on all exogenous variables in the system. The model has a good fit driven by the high correlation between enforcement and collection rates and subsidies and collection rate. The effect of enforce- ment on revenues is shown in table C.4. TABLE C.4: EFFECT OF COLLECTION RATE ON REVENUES 1n R 1n R 1n Coll enf = 1n Coll * enf = 1 (0 + 2 1 enf) = 0.7168 * (2.5719- 2 * 1.2132 * 0.4968) = 0.9795 The elasticity of revenue with respect to enforcement is shown in table Error! Reference source not found.C.5. The estimated elasticity of 0.4866 is very close to the price elasticity of rev- enue and falls within the 95% confidence interval for this parameter. Therefore we cannot con- clude that a price instrument is better than enforcement to increase revenues, or vice versa. The interesting point of this exercise is that the coefficients on enforcement and enforcement squared suggest that collection rates increase at a decreasing rate with enforcement (see the bottom regres- sion in table C.3). So increasing enforcement when collection rates are low may yield higher rev- enues than increasing prices. REVISITING REFORM IN THE ENERGY SECTOR 39 TABLE C.5: EFFECT OF ENFORCEMENT ON REVENUES 1n R 1n R 1n enf= enf enf = 0.9795 * 0.4968 = 0.4866 2 1 , (c) 2004 . 1818 H treet, N.W. Wahington, D.C. 20433, U..A. . : 2003 . , 1 2 3 4 05 04 03 , - , . , , . , . , , . , , , . , - . , , , , . . - , , , . - , , - . Copyright Clearance Center. Copyright Clearance Center, Inc. 222 Roewood Drive Danver, MA 01923, U..A. Tel: 978-750-8400 Fax: 978-750-4470. , Republication Department, Copyright Clearance Center, fax 978-750-4470. , , 202-522-2422. IBN: 0-8213-5689-5 eIBN: 0-8213-5750-6 IN: 1726-5878 - , - , , . - . - . - , , , . - , , , . : , . : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .v . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .vii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 2. . . . . . . . . . . . . . . . . . .9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17 . . . . . . . . . . . . . .17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18 -- . . . . . . . . . . . . . .20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24 4. . . . . . . . . . . . . . .25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28 . . . . . . . . . . . . . . .29 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33 . . . . . . . . . . . . . . . . . . . . . . . . . . .33 . . . . . . . . . . . . . . . . . . . . . . . . . . .35 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41 A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45 iii . 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 . 2. . . . . . . . . . . . . . . . . . . . . . . . . . .10 . 3. . . . . . . . . . . . . . . . . . . . . . .12 . 4. , /. . . . . . . . . . . . . . . . . . . .14 . 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18 . 6. -- . . . . . . . . . .19 . 7. -- . . . . . . . . . . .19 . 8. -- . . . . . . . . . . . . .21 . 9. --, 2002 . . . . . . . . . . . . . . . . . . .22 . 10. -- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 . 11. ( 2002 .) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24 . 12. -- . . . . . . . . . . . . . . .29 . 13. -- . . . . . . . . . . . . . . . . . . . .30 . 14. (/)-- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 1. -- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 2. 2001-2003 . ( . ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34 3. -- ( . ) . . . .35 4. -- . . . . . . . . . . . . . . . .36 5. -- . . . . . . . . .37 6. - -- . . . . . . .38 T .1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42 .1. , . . . . . . . . . . . . . . . . . . . . . . . . . . .46 .2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47 .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48 .4. . . . . . . . . . . . . . . . . . . . . . . . . . .48 .5. . . . . . . . . . . .49 iv , - , , . - , , ; . , - , . - , - , - . , - , - - , . , - . - , , : , . - ; , , , - , . , , v - - , . - - . , - . - , . - , , , , . , - . , , . vii , 2002 . 2003 ., - - ( , ECCU3). -. , - . - , . , , , , , , . USAID, , - , « » . , , 2003 ., . - , . - ( , ECD), : (ECIE), (- ), (), (ECPE) (). (ENV) (DV). (ECIE), (ECIE), (ECHD), (), (ECPE). ix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . « » . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2002 .: 1 . = 2,15 xi , , , - - . , - . , , . - . , - , , , . , . , , . , - . - , - . , , - . , - , /, . 1 2 125 , . , - ( - - ). , - ( 150 ) . . - , . 135%. , , , - ; . , , , . , , . , . , - . ( , ), - . , - , , . , , . - . , , , , - , - . , - . , , - , . - - . , - 3 . - . - , , , . 1 , , - . - , «» , - , , . , , , - - , - . , , , . , - , . , , , - . , (, - ), , , - . - 5 6 , . , . - - , . , , . , . , . - , . , , - , - , , . : (), 1996 . 1; , « » 2002 .2; , - 2000-2002 . , . - . , . ( ) , , , , 2000 2001 . (.1). 3 ( , ). , , , . 1. « : 2000 .» , 2001 . 2. UAID. : . 3. , , . , ; . 7 . 1. 20 16 12 () 8 4 ( ) 0 .2000 2000 .2000 .2001 2001 .2001 .2002 2002 .2002 : , . 2 1990-94 . , 70%. , - . , 2001 . 3,2 . . , 5,4% 2002 . 4 , , , . , .5 1996 . , , ; - - . 1996 . 2002 . 4. No. 19348-GE, « », 1999, No. 22350-GE, « », 2002. 5. , . No. 22350-GE «. », 2002 , . 58, 35. . Ravallion (2001), « : ?» ( No.2665). 9 10 4 %, 18% 2001 . 2002 . 2001 2002 . - , . , ( ) . 2% 1996 2002 ., - (.2). (- ) 8% 10% . - . 1996-2002 . 2%, 20%. 17% (1997-2001 .), 28% 2002 . 7% . - . 90- . - , 98% . - , . - . . 2001 . , 2001 ., . ; - . 2. 80% 60% 40% 20% 0% II IV III II IV III II IV III II IV III II IV III II IV III IV 1996-III 1997-I 1998-I 1999-I 2000-I 2001-I 2002-I : . 11 - 4,5 17 .6 , 2001 2002 . 7,8 , , - . , .9 -- - , . , 1998 . ( 3).10 - - , - , - . , , - . «» , ( ). - , . - , , . - ( - 6. « », , 2002. 7. 39 . 2000 . 164 . 2003 . 300 . . 8. , . , 1998 2000 ., 2000-2001 . 9. , . , . , - , , . 10. . « : .» No. 393. Wahington, D.C. 1998. . 29-30. 12 . 3. 0.16 0.12 0.08 / 0.04 0 1997 1998 1999 2000 2001 2002 : ( ) . , ); (). ( - ), , , . 1999 . (), - . 1998 . , - - .11 , , , -- , - . , , - 1998 ., ; .12 25 .13 2000 . - . 2002 . - 11. . 12. 8 ( 5 % ) . 13. I II. 13 14 - (), - . - ; (UAID), . (- ) ( , , - ). , . , , , - , . - , . . .15 1997 . - ( ).16 - 0,27 /3 0,30 /3 . - 14. , . 15. - , , , . 1998 . , - - , - , . 2002 . (13.7 /), , 10% , 12.4 /. 30 2002 . . , 10% ( , 30%). , . - , . 16. . 4 , , . 14 , ( ) 215 (100 . 2002 .), 17 6,50 18. - , . - , . , - , 1999 ., , , , 10 . 22 . . .19 , , , , - , , (.4). - . , , , , . 4. , /. 80 70 60 50 (1997) 40 () 30 20 () 10 0 -97 -98 -99 -00 -01 -02 : . 17. , . 18. - ( 60 . 1,000 3), . / - , . 19. , - . - , , - . - (. ). , . 15 , , . , , - , . , . ( , ), , , , - . , , . « » , , , , 5 75%. - , . , 2002 . 22 . . , - 20%, 15 /. , , , . , , , . 3 , , ( ) , , , . , «» , . 1996 . , , - , 8%.20 - . , , ( 6,4 8,4%), , - ( ) (.5). 2001 . 2002 . . , , - . 20. 134 . 17 18 . 5. 10% 8% 6% 4% 2% 0% 1997 1998 1999 2000 2001 2002 : . , . 2001 . 94 % - 25% 7% 21. , ( ). , ; , - . , 1997 ., , 1999 . . 1997 ., ( ). - . - , , . , - - , , - , . , - ( ), , . 21. , , . ( 4 ). , . 19 . 6. -- 350 300 250 200 150 . 100 50 0 II IV III II IV III II IV III II IV III II IV III II IV 1996-III 1997-I 1998-I 1999-I 2000-I 2001-I 2002-I : . - 22, , - 200 . /; - 55 ./ (. 6). - , - , (.7). . 7. -- 60% 50% 40% 30% 20% 10% 0% 1997 1998 1999 2000 2001 2002 : . 22. (. ) / , - , (. ). 20 45 51% 1996- 2002 . ( 3% 7% ). , () . , 2 20% - 1999 . , , , - - . - , , . , , , . , , , - . - . , , , . - . - , - , , . , , .23 , , , , . - , , . , , , . -- - . , . 125 (. 8), - 23. , -, , 60-70 , 2-3 . 21 . 8. -- 200 150 100 / 50 0 -00 -00 -00 -01 -01 -01 -02 -02 -02 : . 113 .24 , , - ( ), . . -, - . - 125 - . - , .25 -, , , , , . , , 24. 2000 ., . , 1999 . , . , . - . - - , . 25. ( 5-15 ) 95 /, 3 30 /. 22 . 9. --, 2002 . 7% 6% 5% 4% 3% 2% 1% 0% 125 375 625 875 1125 1375 1625 1875 2125 2375 2625 2875 3125 3375 3625 3875 / : . - . . , , , ; , . , - 875 1,750 (. 9). , 1,500 . , , - , - ( ). , , , . , . - - . (, ) ( ). , - (. 10) - . - . ( , ) 23 . 10. -- 100% 80% (, , ) 60% 40% (, ) 20% 0% III III III III III III 1997-I 1998-I 1999-I 2000-I 2001-I 2002-I : . , , (. - ). , .26 - . , 2001 . 80% . - . 11 - . . , - . , , , , - ( ), . - , , , , - , , . 26. 2002 . 53 % , , 76 % - , . , , , - , . 24 . 11. ( 2002 .) 20 18 1 - 2 16 2 14 12 - 1 i 10 - 8 6 /. 4 2 0 0 20 40 60 80 100 120 . / : « ». , , . , . - . - - . . , , . 125 , . - , - . ( 150 /) . 4 .27 , , , 28 , , . , , , - , , - . - , 27. , - , , . ; , , - , 2002 . 81%, 55% - 93% , . , . , , 2002 . 25%. 28. 40% , 2002 . 25 26 . , . , , , - .29 ( ) , . , - . , - , , , , . . , , , . 2002 . - . 2000-2002 . 91%, 41% 2001-2002 .30 , , . - . , 44 % 2000 . 86% 2002 . 100% , . , - , . , , (. 1).31 . ( - ), , ( , ) . 29. . 30. 1,349 , . - 300 . . , , - . 31. . 27 1. -- 2000 2001 2002 '01 ­ '00 '02 ­ '01 --. 2,79 2,38 1,20 -15% -6% -- . 3,23 2,76 1,29 -14% -20% 86% 86% 93% 0 7 (/) 0,093 0,100 0,124 8% 24% 38% 69% 76% 32 7 --. 2,35 2,31 2,49 -2% 24% --. 217 232 309 7% 33% --. 96 186 266 93% 44% --. 35 44 55 25% 26% 29 37 47 28% 27% 6 7 8 11% 21% --. 61 142 211 132% 49% 44% 80% 86% 36 6 : , . 2002 . . = : . 2001 . . - , , , . , , , , . , - . , , , , . . . - - . (. . 1) , - 28 - . 2001 2002 . 44%, 24%. - 91% 8%. . 32 29 % 2000 . 18% 2001 2002 . - 3 6%. - , , ( , ). - - . , . , - - , , , . , , , , . , , , , , . - , , « , .» - , , , , . , - . - , . - , - . , 32. - . - . - . 2000 . 850 , 2001 2002 . 1000 ; 2003 . 480 . - , . . . 29 , , ; -, , . - , , - . - - - , , - , . . 33 , . (. 12). .34 - . , - , - ( ) ( - . 12. -- 175% 150% 125% 100% 75% 50% 25% 0% -00 -00 -00 -01 -01 -01 -02 -02 -02 : AEC-. 33. , , . 34. , , . 30 ). , - . - , . , , , , ( - , « »), , , , , . , . , 10% - . , ( ) , - . , , , , . - , . , (.13). , , . - , . . 13. -- 175% 150% 125% 100% 75% 50% 25% 0% -00 -00 -00 -00 -00 -00 -01 -01 -01 -01 -01 -01 -02 -02 -02 -02 -02 -02 : AEC-. 31 , . 135 %. - , - , . , , - , . , , , . 5 -- . - . , , . - . - . , , , , . , . - , . , 1994 . , - . 1995 2000 . ( ) 12% 19%. , , 33 34 , . - , 1991 1999 . , . - 287 . , , - , 354 . . , - (, , KfW, ) 202 . .35 , , 2001 2003 . 210 . ( 2). , , . - - , . , , - - . 2002 . ( ) 444 . . 43 . 2001 . 98 . 2003 ., 7,3% - (. 2). (22 . ) 2. 2001-2003 . ( . ) 2003 2001 2002 () 3 000 13 000 36 500 , 6 555 13 646 14 016 , (, ) 21 924 27 346 29 348 , 6 000 10 160 4 500 2 800 3,000 11 500 42 280 69 154 97 867 906 314 1 031 259 1 343 700 % 4,7 6,7 7,3 - 17 279 34 325 46 500 : . 35. . 35 , . . , . 2001 2003 . . . -, . -, - , . -, . - , . 1999 . , 10 . 2002 . ( 3). - 10 , . - 1999 . . 1996 . 10 . , 1998 . 170 . , , , , .36 , . . 35 70 ( - 240 120 ). 3. -- ( . ) 1998 1999 2000 2001 2002 50 15 340 842 10 113 495 0 0 0 0 545 15 340 842 10 113 107 916 123 989 127 190 154 760 182 686 % 0,5% 0,0% 0,3% 0,5% 5,5% : . . 36. . - , ( ). 36 ( ) . - 850 3 .37 , ---- .38 - - ( ). , ( - ) ( 4). , - , - . 2000 2001 . . , , . 27-32% ( 5), 54-64% . - , , , . , - 4. -- - - , 2000 12% 12% 15% 13% 13% 2001 10% 16% 18% 11% 10% , 2000 25% 16% 18% 17% 10% 2001 27% 23% 21% 19% 14% : . 37. , . 38. , - . , , . 2000 . 850 , 2001 2002 . 1000 . 2003 . 480 . . 37 5. -- () % , % , () () 2000 1851 28% 1440 54% 2001 1659 32% 1461 64% 2002 1948 27% 1691 56% : . ( ) , . , - (, ). - , , , , (. 14). , 845 1750 73 145 . , , , . (. 6). . , . . . 14. (/)-- 8% 6% 4% 2% 0% 125 375 625 875 1125 1375 1625 1875 2125 2375 2625 2875 3125 3375 3625 3875 / : . 38 6. - -- - ( %) : 10 16 18 11 10 27 23 21 19 14 (a) 44 38 40 42 39 -- (b) 40 35 43 34 35 (/) 610 561 548 646 535 1000 1000 1000 1000 1000 (a) 407 411 497 476 324 -- (b) 398 384 479 382 287 - (/) 76 70 68 80 66 124 124 124 124 124 (a) 50 51 62 59 40 -- (b) 49 48 59 47 36 : (a) , 875 1750 /. , 125 . 0.124 /. (b) , (a), , , . : . - ( - ) . , . -, , . , - , , 75 125 . . . -, - , 39 ; ( ) (, , , , ). -, . , . , . , . , , , . . (, , , ) - - . . ( , ). , ( , - ) , . (, , ), - . - . , , , , , , «». , , , . . , , , . , , , , . , - , , 40 , - , . - , . - , , , . - - . . , . , , , , . , , . , - , . , - . , - . . .1. 2002 . ; . ( 3) ( 4) 106, . . , , . 6 . . , , , , . 7 , , , . 41 42 T .1. - - - (- , [/. . - . . . 2002() . .]() () )() () . () [1] [2] [3] [4] [5]=10-6[3]/[4] [6] [7]=[5]*[6] [8] 3 0,270 3 412 7,65 70% 10,93 $ 5,08 0,137 35 300 40,15 90% 44,61 $20,75 0,790 32 934 24,04 40% 60,09 $27,95 1,400 42 854 32,67 70% 46,67 $21,71 3 22,563 7 165 200 3.,15 20% 15,74 $ 7,32 : . ( . ), . UAID/ave the Children. . . . . . O 2002 .: 1 . = 2,15 . : . (, ) , . « » . , - . . - , . , , 60 43 44 , , , , 18 , 18 , 60 60 , ( 1 - ). (, , ). , , , - - . 1 , . 2000-2001 . - 8% , - 39 32 %, - 7%, 44%. , - ( ), - , , . 2000 . , , - , - . . 39. , . , . : - , , . , -- . ( - ) . 45 46 , , . , . . , . . (, ) . , . , 2001 . - . . . , , . , - . - , , , . - 15 , . , , . - . , . - , , . (. 1); - , , .1. , - - - 2000-I 13,9 5,3 9,6 4,5 7,1 2,3 2000-II 13,9 4,6 11,1 4,8 7,4 3,1 2000-III 23,1 5,9 10,9 4,6 9,4 3,0 2000-IV 20,7 6,6 9,9 5,3 9,6 3,6 2001-I 11,4 7,9 6,9 4,4 4,3 2,0 2001-II 12,3 8,3 7,5 5,6 6,9 3,4 2001-III 22,6 9,4 9,6 4,8 9,1 4,0 2001-IV 23,5 14,2 12,7 5,8 8,6 4,0 : . 47 . - . , . , , . , - . 2001 ., ( .2): - TATA (. .3). . , 36 . , (R2 F-). , , . 1 2, . , OL - 1.40 , , , . - . 1% 0.7168 % 1% - 0.6565%. - . , 1% 0.2555 %. , , , .2. ln R = 0 + 1 ln Coll + 2 ln P + 3 Temp + 4 Winter01 + 5 ln Subsidy + 1 [1] ln Coll = 0 + 0 enf + 1 enf2 + 2 quality + 2 [2] ln R ­ ln Coll ­ ln P ­ Temp ­ Winter01 ­ 1, 2001 ln Subsidy ­ enf ­ enf2 ­ quality ­ i ­ 40. 2 6.73. , OL 1 , 1 . Tet: Ho: difference in coefficient not ytematic chi2( 1) = (b-B) ' [(V_b-V_B)^(-1)](b-B) = 6.73 Prob>chi2 = 0.0095 48 .3. Instrumental variables (2SLS) regression Source SS df MS Number of obs = 36 F( 5, 30) = 82.64 Model 7.01960499 5 1.403921 Prob > F = 0.0000 Residual .517717153 30 .017257238 R-squared = 0.9313 Adj R-squared = 0.9199 Total 7.53732214 35 .215352061 Root MSE = .13137 ln Revenue Coef. Std. Err. t P>|t| [95% Conf. Interval] ln Coll .7167994 .0784598 9.14 0.000 .5565631 .8770357 ln P .656545 .3602119 1.82 0.078 -.079106 1.392196 Temp_avg -.0030601 .0031612 -0.97 0.341 -.0095162 .0033959 Winter01 -.1818197 .0964468 -1.89 0.069 -.3787904 .015151 ln Subsidy .2666331 .0260078 10.25 0.000 .2135182 .3197481 Constant 9.417039 1.025128 9.19 0.000 7.323448 11.51063 () . reg lncoll enf enf2 quality Source SS df MS Number of obs = 36 F( 3, 32) = 93.61 Model 7.84100824 3 2.61366941 Prob > F = 0.0000 Residual .89351395 32 .027922311 R-squared = 0.8977 Adj R-squared = 0.8881 Total 8.73452219 35 .249557777 Root MSE = .1671 lncoll Coef. Std. Err. t P>|t| [95% Conf. Interval] enf 2.571911 .5743944 4.48 0.000 1.401907 3.741914 enf2 -1.213233 .7249909 -1.67 0.104 -2.689991 .2635254 quality 1547864 .2499061 0.62 0.540 -.3542556 .6638284 _cons -1.766414 .2681718 -6.59 0.000 -2.312662 -1.220165 , , . 1 2. 3. - . , . .4. 49 .4. 1n R 1n R 1n Coll enf = 1n Coll * enf = 1 (0 + 2 1 enf) = 0.7168 * (2.5719- 2 * 1.2132 * 0.4968) = 0.9795 . .5. 0.4866 95% . , . , - , (. . .3) , - , . .5. 1n R 1n R 1n enf= enf enf = 0.9795 * 0.4968 = 0.4866 Revisiting Reform: Lessons from Georgia 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 stimulate public discussion. This paper reviews the changes in the supply of electricity and gas from the perspective of households, utility operators, and the government. The objective is to highlight les- sons from the reforms implemented and to apply them to the future program planned for the rest of the energy sector. The paper concludes that improved service quality and the increased supply of clean and subsidized natural gas have offset the potentially negative impact of higher electricity prices. Despite very good performance by the privatized electricity distribution company in Tbilisi, the sustainability of the reform program is still in doubt. Consolidated government expenditures on energy have increased, but to a large extent this simply recognizes costs that were incurred, but not paid, prior to reform. Existing subsidies to households for electricity provide compensation beyond levels that produce large welfare gains. One option to improve subsidy targeting is to base targeting on actual levels of electricity consumption while providing enough compensation to ensure that the household receives a basic level of electricity. ISBN 0-8213-5689-5 TMxHSKIMBy356890zv":&:<:>:* THE WORLD BANK 1818 H Street, NW Washington, DC 20433 USA Telephone: 202 473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org