39250 Caucasian The Tiger S U S TA I N I N G E C O N O M I C G R O W T H I N A R M E N I A Saumya Mitra Douglas Andrew Gohar Gyulumyan Paul Holden Bart Kaminski Yevgeny Kuznetsov EkaterineVashakmadze The Caucasian Tiger Sustaining Economic Growth in Armenia The Caucasian Tiger Sustaining Economic Growth in Armenia Saumya Mitra Douglas Andrew Gohar Gyulumyan Paul Holden Bart Kaminski Yevgeny Kuznetsov Ekaterine Vashakmadze © 2007 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone 202-473-1000 Internet www.worldbank.org E-mail feedback@worldbank.org All rights reserved. 1 2 3 4 5 10 09 08 07 The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the Board of Executive Direc- tors of the World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The World Bank encourages dissemination of its work and will normally grant per- mission promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750- 4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, World Bank, 1818 H Street, NW, Washing- ton, DC 20433, USA; fax 202-522-2422; e-mail pubrights@worldbank.org. Cover design: Drew Fasick, Serif Design Group Library of Congress Cataloging-in-Publication Data The Caucasian tiger : sustaining economic growth in Armenia / Saumya Mitra . . . [et al.]. p. cm. Includes bibliographical references and index. ISBN-13: 978-0-8213-6811-4 ISBN-10: 0-8213-6811-7 ISBN-13: 978-0-8213-6812-1 (electronic) 1. Armenia (Republic)--Economic conditions--1991- 2. Armenia (Republic)--Economic policy--1991- 3. Investments--Armenia (Republic) 4. Industrial policy--Armenia (Republic) I. Mitra, Saumya, 1951- HC415.17.C28 2007 330.94756--dc22 2007007772 Contents Preface xvii Introduction xix Acronyms and Abbreviations xxxi Part I Policies to Sustain Economic Growth 1 Chapter 1 The Pattern of Growth 3 Chapter 2 Constraints to Sustained Growth 33 Chapter 3 The Framework for Competition 47 Chapter 4 Finance as a Barrier to Accumulation 75 Chapter 5 Impediments to International Integration 93 Chapter 6 Knowledge and Innovation 117 Part II Detailed Analysis 145 Chapter 7 Growth Analysis from the Perspective of Employment Generation and Poverty Reduction 147 Chapter 8 A Social Accounting Matrix for Armenia 193 Chapter 9 Taxation and Economic Efficiency in Armenia 243 Chapter 10 Strengthening Competition 283 Chapter 11 Real and Formal Ease of Doing Business in Armenia in Comparative Perspective: Implications for Regulatory Reforms 329 vi Contents Chapter 12 Armenia's External Performance and Policy Remedies 349 Chapter 13 Civil Aviation Policy 413 Chapter 14 Moving toward Knowledge-Based Competitiveness 489 References 581 Tables Table 1 Key Policy Recommendations xxvi Table 1.1 Armenia's External Debt Burden Indicators 17 Table 1.2 Armenia's Poverty Indicators, 1999­2004 21 Table 1.3 Armenia: Growth and Inequality Decomposition of Changes in Poverty Incidence between 1998­99 and 2004 24 Table 2.1 Progress with Second-Generation Reform, 1994­2004 43 Table 3.1 Selected Indicators of the Quality of Governance, 1998­2004 48 Table 3.2 Values of CPIs for Selected Countries and Groups, 1999­2000 and 2002­04 51 Table 3.3 Relative Ease of Doing Business in CIS and CEEC-10 Economies, 2005 53 Table 3.4 The Size of the Informal Economy and the "Revealed" Ease of Doing Business 56 Table 3.5 Doing Business in Armenia and Selected Comparators in 2004 58 Table 4.1 Financial Indicators in Selected Countries 76 Table 4.2 Bank Concentration Indicator, 2000­04 77 Table 4.3 Decomposition of Interest Rate Spreads, 1999­2004 78 Table 5.1 Average Applied and Bond MFN Tariff Rates in Selected Countries 96 Table 5.2 Telecommunications Sector Overview 104 Table 5.3 Percentage of Households with Access to a Telephone 105 Table 5.4 Basic Telecommunications Infrastructure Benchmarks 106 Table 5.5 Internet Benchmarks 107 Table 5.6 Phased Civil Aviation Reform 113 Table 6.1 The Size of the Shadow Economy 124 Table 6.2 Sequencing of the Armenia KE Policy Agenda 131 Contents vii Table 6.3 Specific Policy Initiatives as Entry Points to Address Systemic Constraints 135 Table 6.4 Matching Policies to Capabilities: A Range of Instruments to Support Innovation 137 Table 7.1 Sources of GDP Growth in Armenia by Sector 148 Table 7.2 Sources of Exports in Armenia, 2002, by Sector 151 Table 7.3 Armenia: Revealed Comparative Advantage, 2003 152 Table 7.4 Direct and Indirect Multipliers 166 Table 7.5 Armenian Economy, Sectoral Structure Depicted by SAM 168 Table 7.6 Sources of Employment in Armenia: Distribution by Sector, 2002 170 Table 7.7 Tables Generated Using the 2002 SAM-based Model for Armenia 176 Table 7.8 Simulation Assuming Export Diversification 178 Table 7.9 An Alternate Approach That Does Not Rely on Export Growth 179 Table 7A.1.1 Effect of a (­50%) Shock on Diamonds 183 Table 7A.1.2 Effect of a (­30%) Shock on Construction 183 Table 7A.1.3 Effect of a (­5.5%) Shock on Remittances 184 Table 7A.1.4 Effect of a (+US$135 million) Combined Shock Based on Relative Export Shares 184 Table 7A.1.5 Effect of a (+10% = US$91 million) Positive Shock in Agriculture 184 Table 7A.1.6 Effect of a (+10% = US$241 million) Positive Shock in Manufacturing and Industry, except Minerals and Mining 185 Table 7A.3.1 Selected Labor Indicator Definitions 191 Table 8.1 Aggregate Social Accounts for Armenia, 2002 196 Table 8.2 Replication of Selected National Accounts Measures by the Macro SAM 197 Table 8.3 Structure of the Disaggregated (micro) SAM for Armenia, 2002 198 Table 8.4 Detailed Sector Descriptions for Each Industry in the 2002 Armenian micro SAM 199 Table 8.5 Armenian Production Statistics by Sector, 2002 202 Table 8.6 Value Added in Production for Armenia, 2002 203 Table 8.7 Tax Payments by Sector in Armenia, 2002 204 Table 8.8 Total Commodity Supply and Component Demand by Consumer, 2002 206 Table 8.9 Household Income by Factor, 2002 209 Table 8.10 Final Demand Disaggregated by Household Type and Commodity, 2002 210 viii Contents Table 8.11 Consumption and Income for Rural Household Decile No. 4 213 Table 8A.1.1 Input Tables Required for the Structured Approach to SAM Development 223 Table 8A.1.2 Input Table #1: Sectoral Production and Trade Values for Armenia, 2002 224 Table 8A.1.3 Input Table #2: 2002 Collected Taxes by Sector and Tax Stream 225 Table 8A.1.4 Input Table #3: Production Technology Structure for Armenia 226 Table 8A.1.5 Input Table #4: Value Added and Consumption Shares 227 Table 8A.1.6 Symbol Table 228 Table 8A.1.7 Mapping GTAP V5 Sectors onto 2002 Armenian SAM Sectors 232 Table 8A.1.8 Mapping of Armenian Output Statistics to SAM Sectors 236 Table 8A.1.9 Mapping GTAP V5 Sectors onto 2002 Armenian SAM Sectors 237 Table 9.1 Economic Sectors in the Armenian Model 255 Table 9.2 Base-Year Production and Trade Statistics by Sector for Armenia, 2002 256 Table 9.3 Benchmark Import Statistics for Armenia, 2002 257 Table 9.4 Export Statistics for Armenia, 2002 258 Table 9.5 Sectoral Value Added, Ranked by Labor Intensity 259 Table 9.6 Estimated Level of Underground and Informal Activity 260 Table 9.7 Armenian Tax Collections, 2002 (by source) 262 Table 9.8 Benchmark Tax Collections, 2002 by Production Sector 263 Table 9.9 Cost of Raising 0.5% GDP (US$15 million) in Tax Revenues: A Comparison of Tax Bases 264 Table 9.10 Indices of Informality and the Average Cost of Funds 265 Table 9.11 Revenue and Welfare Impacts of Selected Tax Reforms 266 Table 9.12 Sensitivity Analysis for the Cost of Raising 0.5% of GDP 268 Table 9A.1.1 Origin of Each Variable 274 Table 9A.1.2 External Data Sources 275 Table 9A.1.3 Production Technology Structure for Armenia 276 Table 10.1 Business Formation Indicators, 2001­04 285 Table 10.2 Employment Indicators, 2001­04 286 Contents ix Table 10.3 Average Earnings Indicators, 2001­04 286 Table 10.4 Gross Value Added Indicators, 2001­04 287 Table 10.5 Foreign Exchange Rate Comparisons, 2001­04 288 Table 10.6 Percentage Share of Different Modes of Transportation, 1999­2003 295 Table 10.7 Internal Freight in Volumes and Percentage Shares, 1999­2003 296 Table 10.8 Welfare Calculations in the Petroleum Market, 1999­2001 302 Table 10.9 Sugar Market Data, 1999­2002 304 Table 11.1 Ranking of CIS and CEEC-10 Economies by Ease of Doing Business in 2005 333 Table 11.2 Size of the Informal Economy and "Revealed" Ease of Doing Business 337 Table 11.3 Doing Business in Armenia and Selected Comparators, 2004 340 Table 12.1 CIS: Total Exports, Export Growth, and Share of Exports in GDP, 1999­2003 353 Table 12.2 Trade in Services by Mode of Supply, 1999­2003 358 Table 12.3 Recipients of Armenia's Exports, 2000­04 364 Table 12.4 Sector Concentration of Exports in Selected Transition Economies, 2003 365 Table 12.5 Factor Intensities of Armenian Trade in 1997, 2000, and 2003 366 Table 12.6 FDI Inflows Compared, 1995­2003 370 Table 12.7 Receipt of FDI Inflows by Sector, 1998­2003 373 Table 12.8 Progress of Structural Reform in Armenia and Lithuania, 1994­2004 378 Table 12.9 Correlation of FDI (1990­2003) and SRI (1999­2003) 382 Table 12.10 Formal Barriers to Doing Business in Armenia, 2003 and 2005 386 Table 12.11 Average Applied and Bound MFN Tariff Rates in Selected Countries 388 Table 12.12 Indicators of Telecommunications Penetration in Armenia and Other CIS Countries, 1998­2003 394 Table 12.13 Ratio of Transportation Costs to Trade, 1998­2003 397 Table 13.1 Zvartnots' Airport Freight Volumes, 1997­2004 418 Table 13.2 Comparative Yerevan-EU Air Fares 421 Table 13.3 Passenger Throughput, Riga Airport, 1990­2005 465 Table 13.4 Czech Airline Data, 1999­2003 467 x Contents Table 13.5 Weekly Frequencies of Air Services Offered 471 Table 13.6 Market Shares 472 Table 13A.1.1 Key Findings, LHR Only 483 Table 13A.1.2 Key Findings, 8 EU Cities 483 Table 14.1 Sequencing of the Armenia KE Policy Agenda 493 Table 14.2 Specific Policy Initiatives as Entry Points to Address Systemic Constraints 495 Table 14.3 Gross Domestic Product by Sector, 1998­2003 500 Table 14.4 Sector Contributions to Export Growth, 2000­02 500 Table 14.5 Countries That Have Successfully Aligned Key Drivers of Growth and Developed Innovation Clusters 503 Table 14.6 Countries with Obstacles to Aligning Key Drivers of Growth 504 Table 14.7 Exports of Goods and Services, 1995­2002 515 Table 14.8 The Size of the Shadow Economy 515 Table 14.9 Key Innovation Indicators, 1997­2002 519 Table 14.10 Interest Rate Spread, 1996­2002: Armenia and Comparators 520 Table 14.11 Sequencing of the Armenia KE Policy Agenda 528 Table 14.12 Specific Policy Initiatives as Entry Points to Address Systemic Constraints 530 Table 14.13 Matching Policies to Capabilities: A Range of Instruments to Support Innovation 533 Table 14.14 FDI Incentives in Armenia versus Comparator Countries 557 Table 14A.1.1 Annual Growth Rates of Total Factor Productivity 564 Table 14A.3.1 Gross Capital Formation, 1995­2002 569 Table 14A.3.2 Performance--Economic Regime Variables 570 Table 14A.3.3 Governance Variables 571 Table 14A.3.4 Innovation Systems Variables 572 Table 14A.3.5 Education Variables 573 Table 14A.3.6 ICT Variables 574 Figures Figure 1.1 Official Development Assistance, Cumulative, 1992­2003 4 Figure 1.2 ECA Countries, Income per Capita and Employment in Agriculture, 2003 5 Figure 1.3 Index of Real GDP, 1990­2005 7 Figure 1.4 GDP Growth Rate, 1996­2006 7 Contents xi Figure 1.5 GDP Growth and Contribution of Expenditure to GDP, 1996­2005 8 Figure 1.6 Evolution of REER, Trade Balance, and Current Account Balance, 1999­2005 8 Figure 1.7 Relative Productivity Level and Productivity Growth Rate, Latest Data 9 Figure 1.8 Relative Unit Labor Cost Index, 1997­2004 10 Figure 1.9 REER Quarterly Increase, Indicating Appreciation 12 Figure 1.10 Key Economic Indicators of External Performance, 1998­2005 13 Figure 1.11 Savings and Investment Balance, 1994­2005 14 Figure 1.12 Gross Domestic Savings and Gross National Savings, 2004 15 Figure 1.13 Sectoral Shares in GDP and Investment by Sectors, Average, 1999­2005 16 Figure 1.14 Indicators of Public and Publicly Guaranteed External Debt under Alternative Scenarios, 2005­25 18 Figure 1.15 Growth Incidence Curves, 1998/99­2004 22 Figure 1.16 Participation, Employment, and Unemployment Rates in Urban and Rural Areas (population aged 15 and older) 26 Figure 1.17 Change in Sector Prices, Nominal Value- Added Deflator 27 Figure 1.18 Nonagricultural Jobs in the Productive State and Private Sectors, 1998­2003 28 Figure 1.19 Relative Labor Productivity by Sector 30 Figure 2.1 Progress in First-Generation Reforms as Revealed in Values of the EBRD-Based Aggregate Index in 1992­2004 in Selected Transition Economies 38 Figure 2.2 Challenge of Sustaining and Diversifying Sources of Growth toward Those with Stronger Spillover Effects 39 Figure 2.3 Progress in Structural Reforms: Values of SRI, 1994­2004 41 Figure 3.1 Aggregate Quality of Governance, 1998­2004 49 Figure 3.2 The Average of Government Effectiveness and Regulatory Quality, 1998­2004 50 Figure 3.3 Gasoline Price Structure for the Period 1999­2001 68 Figure 5.1 Armenian Exports, 1997­2004 94 xii Contents Figure 6.1 Armenia and the World: Knowledge Economy Index 119 Figure 6.2 Armenia: Performance in the Four Knowledge Economy Pillars, 1995­Most Recent 121 Figure 6.3 Armenia Knowledge Economy Index Scorecard 121 Figure 6.4 Economic Incentive Regime: Armenia and the World 122 Figure 6.5 Armenia's Scorecard in the Economic Incentive Regime 123 Figure 6.6 Education: Armenia and the World 125 Figure 6.7 Armenia's Scorecard on Education 125 Figure 6.8 Innovation: Armenia and the World 126 Figure 6.9 Armenia's Scorecard in Innovation 127 Figure 6.10 ICT: Armenia and the World 127 Figure 6.11 Armenia's Scorecard in ICT Variables 128 Figure 6.12 Virtuous Circle of Growth and Reforms 133 Figure 6.13 The Pyramid of Learning Capabilities of Firms 136 Figure 7.1 Armenia and the Slovak Republic, Employment by Sectors, 1994­2004 148 Figure 7.2 Sectoral Shares in GDP and Investments 149 Figure 7.3 Matrix: Classification of Armenian Exports by Composition and Destination, 2003 153 Figure 7.4 Real Growth and Factors of Production, 1997­2004 157 Figure 7.5 Real Wage Growth and Productivity Growth, Unit Labor Cost, 1998­2004 158 Figure 7.6 Labor Productivity by Sector 159 Figure 8.1 Expansion from a Single RA to Multiple Households 208 Figure 8.2 Impact of Increased Wages or Lower Food Prices on Poverty for Rural Decile No. 4 214 Figure 8A.1 Simplified Structure of SAM Development Process 223 Figure 9.1 Armenian Production Structure for Formal and Informal Activities 249 Figure 9.2 Demand for Formal and Informal Goods 252 Figure 9A.1.1 Armenian Agricultural Accounts for 2002 278 Figure 9A.1.2 Industrial Production Shares According to the National Statistical Service 279 Figure 9A.1.3 Official Armenian National Production Statistics for 2002 280 Figure 10.1 Business Competitiveness Indicators for Selected Sectors 287 Contents xiii Figure 10.2 Gasoline Price Structure, 1999­2001 303 Figure 10.3 Sugar Price Decomposition 303 Figure 10.4 Interest Rate Spreads 310 Figure 10.5 Domestic Credit to the Private Sector 311 Figure 12.1 Dynamics of Exports of Goods and Services and GDP and Change in Exports and Imports as Percent of GDP, 1998­2003 355 Figure 12.2 Exports of Diamonds and Other Products, 1997­2004 361 Figure 12.3 Privatization and FDI Inflows, 1998­2004 372 Figure 12.4 Progress in Structural Reforms: Values of SRI, 1994­2004 377 Figure 12.5 Total FDI Inflows per Capita and SRI, 1999­2003 381 Figure 12.6 Quality of Infrastructure in Armenia against Georgia, the CEEC-8, and Highly Developed Countries 391 Figure 13.1 Zvartnots' Airport Passenger Volumes, 1990­2005 419 Figure 13A.1.1 Average Fare 485 Figure 13A.1.2 Yield by Distance 485 Figure 14.1 Armenia and the World: KEI 509 Figure 14.2 Armenia and Comparators: Disaggregated Performance in the KE 511 Figure 14.3 Armenia: Performance in the Four KE Pillars 511 Figure 14.4 Armenia KE Index Variables, Basic Scorecard 512 Figure 14.5 Economic Incentive Regime: Armenia and the World 513 Figure 14.6 Armenia's Scorecard in the EIR 514 Figure 14.7 Education: Armenia and the World 516 Figure 14.8 Armenia's Scorecard on Education 517 Figure 14.9 Innovation: Armenia and the World 517 Figure 14.10 Armenia's Scorecard on Innovation 518 Figure 14.11 ICT: Armenia and the World 520 Figure 14.12 Armenia's Scorecard on ICT Variables 521 Figure 14.13 Armenia: Real GDP per Capita: Alternative Projections, 2003­20 522 Figure 14.14 Virtuous Circle of Growth and Reforms 529 Figure 14.15 The Pyramid of the Learning Capabilities of Firms 532 Figure 14.16 Comparative Performance of Armenia's ICT Infrastructure 535 Figure 14A.4.1 Armenia and ECA 575 xiv Contents Figure 14A.4.2 Armenia and Latvia 575 Figure 14A.4.3 Armenia and Slovenia 576 Figure 14A.4.4 Armenia and Russia 576 Figure 14A.4.5 Armenia and Israel 577 Figure 14A.4.6 Armenia and Chile 577 Figure 14A.4.7 Armenia and Costa Rica 578 Figure 14A.4.8 Armenia and Ireland 578 Boxes Box 2.1 Armenia--Production Statistics by Sector 36 Box 2.2 Gains from the Liberalization of the Services Sectors 44 Box 3.1 The Petroleum Distribution Market 69 Box 4.1 Priorities in Reform of Financial and Capital Markets 80 Box 6.1 Korea's Transition to a KE: Bottom-up Initiative Leads to Government Action 130 Box 6.2 Distance Learning as a Potential Pilot Project to Enhance Education-Industry Linkages 142 Box 7.1 Reconciliation of National Accounts and the Household Survey Data 155 Box 7.2 Linkages in the Model Based on the Social Accounting Matrix 163 Box 7.3 Information and Communications Technology Sector 180 Box 10.1 Noncompetitive Behavior in the Gasoline Market 301 Box 10.2 Anti-Competitive Practices of the Customs Office 306 Box 12.1 Lycos Armenia on the Move 359 Box 13.1 Armenian Aviation Market Entry Restrictions 415 Box 13.2 Case Studies: Additional Liberalization Experiences 428 Box 13.3 Economic Benefits of Liberalization: An Illustrative Calculation 429 Box 13A.1.1 Sequence of Data in Excel File 481 Box 13A.1.2 Impact of Ryanair on the Market 484 Box 14.1 The Growing Importance of Knowledge: Global Trends 505 Box 14.2 The Four Pillars of the KE 508 Box 14.3 Armenia' Growth Paradox, Illustrated by Private Companies 523 Contents xv Box 14.4 Korea's Transition to a KE: Bottom-up Initiative Leads to Government Action 526 Box 14.5 Distance Learning as a Potential Pilot Project to Enhance Education-Industry Linkages 539 Box 14.6 Toward a National System of Support for Continuous Vocational Training 540 Box 14.7 The Fundación Chile Model and Its Relevance for Armenia 541 Box 14.8 Clusters and Value Chains 543 Box 14.9 Private Universities in Turkey and the Turkish Diaspora of the Highly Skilled 552 Box 14.10 Toward a New Type of Diaspora Involvement in Armenia's Economy: Lessons of the University Alumni Model 553 Preface This book is the result of work carried out by the World Bank's eco- nomic team charged with providing analysis and policy advice to the Armenian government in the years 2005­6. These were years in which Armenia had established a decade-long solid record of sustained eco- nomic growth amid conditions of economic stability, low inflation, and modest fiscal deficits and external debt. For an economy that had endured a collapse of economic output and hyperinflation in its imme- diate post-Soviet period, this was a signal achievement, particularly when it is noted that the country is poorly endowed with natural resources, and its two major borders have been closed by its neighbors to trade and labor movement since its independence. The attribution for the emergence of Armenia as the Caucasian Tiger lies in the creation of an environment of macroeconomic stability and the determined pursuit of reforms aimed at establishing a market economy that was integrated with the world. This book describes that story. But the central focus of the book remains the reform agenda for the future. It is argued that the continuation of high rates of growth would require building defenses against economic shocks the country may face, and this would entail addressing the key vulnerabilities in today's economy. Thus the book is intended not only as a case study of success in post-Soviet economic transition, but also as a candid piece of policy advice for the Armenian authorities. This book was written by a team headed by Saumya Mitra and consisting of Doug Andrew, Gohar Gyulumyan, Paul Holden, Bart Kaminski, Yevgeny Kuznetsov, and Ekaterine Vashakmadze. It utilizes the contributions of Reza Ghasimi, Karen Grigorian, Armine Khacha- tryan, Artsvi Khachatryan, Matin Kholmatov, Miles Light, Gareth Locksley, Stephen Miller, Thomas Rutherford, Sue Rutledge, Vahe Sahakyan, Eduardo Siandra, and Michel Zarnowiecki. Comments from Shahrokh Fardoust, Felipe Jaramillo, Samuel Otoo, Christian Petersen, Roger Robinson, and Roberto Zagha are gratefully acknowledged. The xvii xviii Preface book has benefited from close cooperation with the IMF team, headed by Hassan Al-Atrash, working on Armenia. It was processed by Zakia Nekaien-Nowrouz. The authors are particularly grateful to Donna Dowsett-Coirolo and Cheryl Gray for their judicious guidance and steadfast support for this work. The authors are also indebted to the Armenian economic team headed by Vardan Khachatryan, minister of finance, for its many valuable observations on the analysis contained in this book. The book draws on sector-specific studies that have been carried out by colleagues at the World Bank; in particular, the study on pov- erty assessment (World Bank 2006), prepared by a team headed by Alexandra Posarac, and the study on the financial sector assessment program (World Bank and International Monetary Fund 2005c), pre- pared by a team headed by Hormoz Aghdaey. The responsibility for the text is entirely that of the authors. Introduction T his book is intended to explain the factors underlying the stellar growth record that has led to Armenia's emergence as the Cauca- sian Tiger and to provide policy advice to the Armenian authorities to ensure the continuation of this growth. The book is presented in two parts, with Part I containing the analysis and the policy advice and Part II containing detailed background papers. In Part I, Chapter 1 sketches the pattern of growth, indicates the opportunities for broad- ening growth and establishes linkages to employment and poverty. Chapter 2 examines the record of reform and the principal constraints to growth in the future, using a growth diagnostics framework being piloted in the World Bank and making use of comparisons with the experiences of more advanced post-transition reformers. Chapters 3­6 investigate the key opportunities for reform: strengthening the framework for competition, lowering the costs of financial interme- diation, developing policies that will deepen the integration of the economy into global markets, and creating the conditions for building on knowledge assets and promoting innovation. GROWTH, POVERTY, AND EMPLOYMENT Over the past half-decade the Armenian economy has grown by double- digit rates annually on average--reminiscent of the east Asian tiger economies--and high growth rates have been registered for over a dec- ade. This impressively consistent performance can be attributed to the steadfast pursuit of market-oriented reforms, assisted by large external inflows on grant or soft terms. Nevertheless, Armenia remains poor, with income per head at around US$2,100 today or about a third of that in the Baltics. Growth has now begun to make a significant dent in poverty, but its impact on employment has been muted. This book will examine the reasons for the weak transmission link to employment and xix xx Introduction will suggest policy reforms. It will discuss the major sources of risk to the continuation of high growth and will outline the policy adaptations necessary to place growth on firmer foundations. To meet the aspira- tions of the public, economic policies will have to be designed to max- imize the prospects for growth; and, thus, impediments to growth will have to be addressed early and vigorously. Armenia was quick to recover from the output shock experienced by transition economies: its upturn in output, dating from 1994, cor- responds to the experience of Central Europe and the Baltic States and predates by four to five years the recovery in the rest of the former Soviet Union (FSU). Growth has been driven by productivity gains in the private sector as macroeconomic stability took hold; the role of private markets was rapidly expanded; the public sector remained small as a share of the total economy; and important institutional measures were adopted to ensure free price formation, liberal trade in goods services and investment, private ownership of assets (including land), and industrial restructuring. Moreover, the adoption of respon- sible fiscal and monetary policies in the late 1990s led to the defeat of inflation and to predictability in the stance of financial policies. Thus, first-generation structural and institutional reforms were achieved and have laid the foundations of the impressive growth performance. The recent period of growth has resulted in a sharp rise in the consumption of the poor. The poorer quintiles of the income distri- bution have gained more from growth than the richer quintiles, with the extremely poor showing the greatest gain. Overall poverty fell from over half to one-third of the population between 1998 and 2004 and extreme poverty fell from over 20 percent to just 6 percent of the population over the same period. The reduction in poverty has been driven by consumption growth rather than by redistribution effects. The capital city of Yerevan has reaped large dividends from the fall in poverty, with other urban areas benefiting to a much lesser degree. Despite the impressive GDP growth, unemployment, according to survey data, stands at around one-fifth of the labor force; and a dual labor market has emerged with large underemployment or subsist- ence employment coexisting with a more skilled labor force that has enjoyed large real wage gains in expanding sectors of the economy. The causes of the weak response of employment to investment and growth lie in a business climate that has discouraged the flexible use of labor and in inadequate skills among the unemployed. Address- ing the binding constraints to the formation and expansion of firms remains central to raising employment. More broadly, this book argues that the incompleteness of structural reforms and inadequacies in the development of institutions and practices that encourage competition are factors leading to the persistence of unemployment. Introduction xxi This book suggests that the continuation of high rates of productiv- ity gains in the economy requires Armenia to make use of new oppor- tunities in reforms directed at sharpening the competition framework of the economy, achieving closer integration with international trade and capital markets, deepening financial markets, and creating the conditions for the absorption of knowledge leading to higher tech- nological sophistication. Should the country fail to seize the oppor- tunities to advance in this second generation of reforms and to over- come the associated institutional and political difficulties, high rates of growth over the next decade would be put at risk. Growth remains narrowly based, despite some widening of the base in recent years, exports remain concentrated by product and markets, significant bar- riers to domestic and private investment remain within a context of a weak competition framework and high policy-induced costs, and the country has made poor use of opportunities for knowledge and innovation. All these factors make up constraints to growth over the medium term. The challenge to policy makers now lies in seizing the opportunities for reform by breaking through these constraints and capitalizing on the solid reform achievements thus far. SEIZING NEW OPPORTUNITIES FOR REFORM As experience has repeatedly shown--not least among the south- east Asian tigers--rapid growth over prolonged periods can conceal underlying weaknesses in an economy and can lull policy makers into complacency. Armenia has begun to address the second generation of reforms somewhat fitfully. Utility reform and privatization have proceeded, and some improvements to the business environment have been made; however, a primitive state of corporate governance, poor property rights enforcement, arbitrary practices in the tax and cus- toms administrations, the distortions associated with corruption, and stunted financial markets all act to impose a large risk premium on investment. Surveys conducted over a number of years and using varying methodologies provide robust evidence of the importance of these factors in inhibiting business. Policy shortcomings in the tele- communications and civil aviation regimes in particular effectively levy a large tax on all economic activities. This environment of weak second-generation reforms has resulted in a dualistic output market (parallel to the dual labor markets), in which entrepreneurs enjoying property rights protection, benefiting from distortions in tax and cus- toms practices, and having recourse to internal financing, gain from being incumbents, while potential entrants to the market are placed at a disadvantage. xxii Introduction The policy adaptations discussed above are also essential to deal- ing with the weak response of employment to high rates of economic growth in Armenia. The initial output and price shocks of transition led to a shift of labor to subsistence agriculture and to the develop- ment of a large informal economy. Macroeconomic stability, privatiza- tion, and enterprise restructuring have promoted flexibility in labor markets, but deficiencies in the business environment, some rigidities in the employment regime, and inadequate financial markets have led to a dual labor market: there is a market of real wages that are rising on the strength of impressive productivity gains, and another market (a larger one) which is informal, stagnant, and immobile. This book can be seen as an attempt to analyze the policy responses necessary to make the growth in rates of employment commensurate with output growth. Investment has been financed to a disproportionate degree from external sources, though the domestic share of investment financing has been rising in recent years. The opportunity offered by a further bout of external grant financing over the medium term, with Armenia qualifying for a large increase in U.S. bilateral assistance in the form of the Millennium Challenge Corporation, can prove a prize if the policy environment is reformed so as to deal with the weaknesses outlined above; or it can be a curse if, in a weak policy setting, private invest- ment is crowded out and excessive consumption or nonsustainable public investment is encouraged. Again, as experience shows, under conditions of grant inflows, the pursuit of reforms can become more difficult and the quality of growth can often be sacrificed. This under- scores the importance of redoubling reform efforts in a farsighted manner in order to position the economy for the time when external support will dwindle. The rest of the book focuses on how to make the best use of this window of plentiful external funds. SHARPENING COMPETITION Within the Armenian economy, competition is hobbled by weaknesses in the law and its application (as with the competition law and the law on contracts) in formal and informal institutions and their workings (such as corporate governance arrangements); by business regulations and governance rules and practices (such as arbitrary and unpredict- able behavior of the tax and customs administrations); and by policy- induced high costs that fall on the shoulders of all enterprises (such as telecommunications and transport costs). The competition law, a recent innovation, needs to be strengthened in both structure and implemen- tation, notably by clarifying what is meant by collusive behavior and Introduction xxiii dominant market practices, and by modernizing the basis for judging pricing practices. Implementation is bedeviled by a weak commission and by the inadequate legal powers of the commission. The commer- cial transactions law offers insufficient safeguards, with the result that contracts tend to be informal. Equally worrisome is the state of corporate governance, where laws, regulations, and institutions are particularly weak. Regarding several key transparency issues--notably public access to company registries and lists of founders and shareholders--Armenia is among the weak- est of the CIS countries. The quality of corporate financial reporting also remains deficient. Moreover, while the securities law has some weaknesses, the greatest corporate governance deficiencies are found in the company law and other basic legislation affecting the corporate sector. Weak corporate governance practices reinforce the dominant position of a limited number of businessmen and undermine investor confidence. Furthermore, informal barriers to competition--such as the system of interlocking obligations arising from favors and inter- ventions among businessmen--act as a substantial barrier to entry for foreign investors as well as for new local businesses. Despite satisfactory indicators in a number of categories measuring the ease of doing business (such as ease of formal entry, hiring employ- ees, and business exit), Armenia falls short in other areas, the weakest of which are the registration of property, access to credit, property rights, and large severance payments (at least until the recent reform of the Labor Code, now being put into effect). Overall, however, even in the most streamlined and reformed areas, the level is well below the best practices in transition economies. In addition, Armenia has a large informal economy, with the main incentives to informality being the arbitrary behavior of public servants, the burden of administrative regulation, and a poor record in establishing the rule of law. Absent in Armenia is a major attraction of formality--the access to credit for which formal registration is a precondition. Such constraints to enterprise activity have implications for effi- ciency and also carry growth costs. The weaknesses in market insti- tutions act as a barrier to the development of financial markets, and, in turn, this leads to the dominance of the incumbent in markets. Financial intermediation is shallow, and the cost of funds is high. DEEPENING FINANCIAL INTERMEDIATION Confidence in banks is affected by the shortcomings in corporate gov- ernance, notably the lack of transparency regarding the ownership and control of banks. Moreover, the opaque ownership structure of xxiv Introduction the corporate sector makes it more difficult for the banking sector to assess risks concerning corporate sector exposures. Official dis- closures of direct and indirect ownership are substantially weaker than in other transition economies. No disclosure of beneficial owners is required under Armenian law. This limits the ability of banks to ensure compliance with prudent limits on loan portfolio concentra- tions and related party transactions. These shortcomings in corpo- rate governance can be addressed through legal, institutional, and supervisory reforms. The Central Bank of Armenia has developed a proposal to accelerate improvements in corporate governance in the banking sector through legislative and supervisory means and by introducing upward-consolidated supervision. The protection of creditor rights is one of the main legal problems hindering the growth of bank lending. The enforcement of collateral is difficult for banks, and this translates into high lending rates. In Armenia, movable property cannot be used effectively as collateral to secure loans. There are numerous problems with the secured transac- tions framework. Floating pledges are not allowed, and there are no registries that allow pledges to be perfected. Repossession is time- consuming and costly. A particular problem with financing exports is that there is no provision that allows the use of future produc- tion as collateral, so financing against export orders is not feasible. A similar problem arises with imports because there is no provision for pledging goods that are not in the possession of the borrower; goods that are being imported, even though they have been paid for by letter of credit, cannot be seized in the event of default. The effect of the inadequacies of the collateral framework extends throughout the economy, with the result that banks correctly perceive lending as extremely risky. Overall, Armenia lacks a regularized system of credit that should be stimulated by mechanisms providing efficient, transparent, and reli- able methods for recovering debt. Armenia's financial sector (possibly with help from the Central Bank and the Ministries of Finance and Justice) should promote an informal, out-of-court process for dealing with cases of enterprises with financial difficulties in which banks have a significant exposure. Finally, given the vital role played by the generation and transmis- sion of information on creditworthiness in ensuring low intermedia- tion costs, it is suggested that the Central Bank adopt an active role in facilitating private credit bureau development by making credit bureau reports a requirement for lending. This requirement should be coupled with mechanisms that ensure efficient entry and exit by credit bureaus, thereby avoiding abuse of their exclusive rights; and, in order to ensure greater privacy protection, there should be a requirement Introduction xxv that borrowers must give explicit consent prior to any access to their files. In addition, improving governance in credit bureaus and raising the quality of information disseminated remain important. REMOVING BARRIERS TO INTERNATIONAL INTEGRATION Armenia enjoys an admirably open regime in trade in goods and services and in capital and investment flows. No legal restrictions are in place for foreign capital inflows; the foreign investment regime provides for national treatment, most favored nation (MFN), and full repatriation of capital and earnings. Access to Armenian mar- kets of goods is liberal in terms of official border and behind-the- border arrangements. Tariffs are low not only by CIS standards but by international standards as well. Although both its weighted and unweighted average MFN tariff rates are twice as high as in the Euro- pean Union (EU), they are still well below 5 percent. Furthermore, commitments made upon accession to the World Trade Organization (WTO) have infused a considerable degree of stability and predict- ability into Armenia's foreign trade policy and have also reduced the potential for the capture of foreign trade decisions by narrow interest groups by providing the government with tools for taming the rent- seekers. Armenia's two-band tariff regime, with applied MFN tariff rates at zero or 10 percent ad valorem, has been locked. Armenia has made highly liberal commitments under the WTO General Agreement on Trade in Services (GATS). Except for telecom- munications, Armenia's bound sectoral commitments are extensive in terms of both coverage and market access across different modes of supply of services. The number of subsectors in which exceptions are placed on a mode of supply (that is, unbound) is very small in Armenia's schedule. Yet neither commitments under the WTO agreements nor legal provisions protecting private property rights and enforcing contracts alone have ensured the high contestability of Armenia's domestic markets. First, the capacity of courts remains weak, as they operate slowly in enforcing contracts and mediating commercial disputes. Sec- ond, the computerization of customs has not to date much improved the quality of customs services, and customs procedures have yet to achieve WTO standards of transparency. The time needed to complete customs clearance is grossly excessive. Value added tax (VAT) reim- bursements do not occur quickly enough and are underpaid by the government, though there was a substantial improvement in 2005. Weaknesses in the provision of backbone services add greatly to the cost of participating in the emerging division of labor. Falling xxvi Introduction transportation and communications costs create opportunities for out- sourcing, just-in-time production, and supply-chain management. The high transportation and communications costs in Armenia are barriers to participation in the division of labor based on production fragmenta- tion. This is illustrated particularly by services in telecommunications and civil air transport. The government's decision to grant a legal monopoly (originally until 2013) to ArmenTel, the local telecommunica- tions company owned by a foreign investor, has led to greatly increased costs and to a considerably lowered quality of service in this key eco- nomic activity. The cost of using the Internet in Armenia was 41 percent higher than the average for the CIS. The prices charged by ArmenTel for high-speed connections are 30 times those of countries with competition in telecommunications. Armenia has a surprisingly low number of Inter- net users, well below the CIS average. The poor quality of the information technology infrastructure is a barrier to the development of the economy, and the very high telecom- munications costs severely exacerbate the disadvantage associated with Armenia's geographical location. ArmenTel's dominant position has been reduced by the introduction of a cellular competitor in 2005, and provision has been made for a second cellular operator in 2009. However, the new mobile operator was licensed in a process that was not transparent and fell far short of international good practice; it is unlikely that a competitive regime will emerge. Duopolies can pro- duce outcomes close to monopolies unless competition law is actively applied to prevent tacit collusion. The analysis of this book suggests that the regulator should be given enhanced powers in an attempt to countervail the adverse effects of ArmenTel's continuing dominant market position. Further- more, future policy should be developed and announced now. This policy should stipulate that all government-imposed economic entry barriers would be removed at the same time that the regulator is given enhanced powers; in addition, the available and necessary radio fre- quency spectrum should be auctioned off well before 2009 so that further mobile operators, to the extent that they consider they would be commercially viable, would be able to enter the market as soon as restrictions are lifted. After 2013, with the termination of ArmenTel's right to exclusivity, the market at that point should be regulated only by competition law. Freight shipped by air from Yerevan fell by more than two-thirds between 1997 and 2005--a dramatic development considering that the value of exports of goods rose 4 times in dollar terms. The fall in the share of the most dynamic worldwide mode of transport--air trans- port--can be attributed to the fall in the competitiveness of air trans- port services, largely because of a restrictive aviation policy. Armenia Introduction xxvii has operated a bilateral system of air transport regulations that is based on a restrictive positive list approach, which limits the provision of services to those services that are explicitly permitted. The empirical evidence from countries that have deregulated the domestic aviation sector (for example, the United States and the EU countries) is robust: passengers and air freight shippers in both the EU and the United States have experienced a dramatic and continu- ing decline in airfares. The costs of bilateral aviation agreements are not confined simply to the higher prices of air transport but include more difficult-to-estimate costs of forgone opportunities. These costs are potentially large, as lower airfares boost tourism, stimulate impor- tant flows of ideas and human capital, and deepen networks. For a small, land-locked economy aiming to maximize its long-term eco- nomic growth rate, the best aviation policy would be to eliminate government-imposed entry barriers to air transport. Armenia could gain considerably from following Open Skies arrangements among CIS countries modeled on the EU European Civil Aviation Area Model as well as joining the many countries that have signed Open Skies agreements with countries such as the United States. Because of the limited contestability of markets and deficiencies in backbone services, information, communications, and technology (ICT) firms have not been able to crest the recent change in the ICT sector worldwide. Armenian firms have successfully entered software and imaging technology niches, but they have failed to enter other stages of the production and delivery processes--in particular as pro- viders of front-end customer contact/support services or suppliers of components. CAPITALIZING ON INNOVATION AND KNOWLEDGE The generation and diffusion of knowledge and the use of innova- tive technology are important factors behind growth, in general, and the raising of labor productivity, in particular. Information technology externalities that promote the expansion of producer services (such as communications, sophisticated financial and insurance products, and marketing) explain a significant portion of the total factor pro- ductivity growth. Armenia has the advantage of a large proportion of highly educated population; however, this educated segment has largely obsolete specialized skills and has eroded through emigration, low public spending on education, and delayed reforms in university education. The country has experienced an entrepreneurial diaspora; and it also suffers from weak local entrepreneurship, in which clus- ters and value chains are not developing. Entrepreneurs and policy xxviii Table 1 Key Policy Recommendations Matching Reform Priorities to Policies and Instruments Reform priorities Policy objectives Instruments and interventions Intr Create Raise the domestic investment · Improve the business climate, simplify regulations and tax code, and create oduction macroeconomic ratio, raise public investment, and measures to fight corruption. policies to foster increase the effectiveness of public · Redirect budget resources to public investment, particularly to enhance human growth. spending. capital and build rural infrastructure. Sharpen Foster competition and encourage · Strengthen the competition law by clarifying collusive behavior and dominant competition. private sector development by market practices and by modernizing the basis for judging pricing practices. lowering cost of doing business. · Enhance the administrative capacity of the competition commission to review cases and enforce its rulings. · Modify the commercial transactions law to strengthen safeguards. Integrate into Create infrastructure that facilitates · Establish a "white list" of firms eligible for special treatment by Customs and VAT global markets and encourages firms to engage in Administration, including quick customs clearance. for goods and international trade. · Use provisions of the WTO Agreement on Customs Valuation instead of using services. reference prices. · Create an effective scheme that provides duty waivers and exemptions from other restrictions on imported inputs. · Give rebates of VAT as soon as exports are cleared by customs. · Extend direct transfer input to all customs houses. · Bring customs-related documents in line with what is really required under a computerized system such as the Automated System for Customs Data (ASYCUDA). · Simplify customs clearance procedures for exports with a strict time limit on releasing a shipment; if it is exceeded, a shipment should be immediately released. · Support participation in international R&D networks (such as the EU 6th Framework Program). · Sign the WTO Information Technology Agreement. Support Liberalize air services markets. · Deepen implementation of liberalised government civil aviation policy. international Maximize the contribution of ICT. · Increase and add new services under "open skies" type arrangements. integration. · Pass the new telecommunications law and pass to the Public Services Regulatory Commission (PSRC) responsibility for implementation, including the ArmenTel license. · Promote cellular competition. · Have the PSRC prepare a regulatory policy statement that, among other things, lays out an implementation plan for the entry of a third service provider by 2009. Deepen financial Strengthen corporate governance · Implement new law on corporate governance of banks. markets. standards and practices. · Strengthen official disclosures of direct and indirect ownership. Build legal foundations for · Make disclosure of beneficial owners required by Armenian law. protection of creditor rights. · Ensure public access to the company registry and lists of founders and shareholders. Improve information flows and · Implement law on creditor rights and secured transactions with streamlining of promote capital markets. judicial procedures. Foster investment activity through · Strengthen supervision of capital markets. expanding access to bank credit to · Implement unified supervision of financial and capital markets, with new standards small and medium firms. for supervision of insurance, pensions, and housing finance institutions. · Promote the development of a credit bureau. Capitalize on Foster competitiveness by · Provide business advisory and support services, such as SME and microenterprise innovation and introducing basic innovation skills support agencies. technology. and encouraging adoption and · Facilitate access to finance (including microfinance). application of new ideas. · Develop management and skills. Intr Support market development and · Develop internet-based information services. oduction entry into global value chains · Support technology adoption and adaptation projects. by fostering strategic alliances · Adopt cluster-based approaches to stimulating innovation. and certain in-house innovation · Support participation in international R&D networks (such as the EU 6th capabilities. Framework Program). Diffuse experience of innovation · Encourage participation of national innovation leaders in national advisory bodies, xxix leaders as role models for the rest technology foresight, and cluster processes. of the economy. xxx Introduction makers alike do not appreciate (and hence do not seek to improve) the value of intangibles (such as brand names, business reputations, marketing and managerial skills, and networks). The policy challenge lies in the mobilization and recombination of the existing human capital, triggered by an initially modest invest- ment in intangibles, such as mechanisms of knowledge and skill trans- fer from the diaspora to Armenia. International experience suggests that, although major reform efforts from the top are vital, bottom-up experiments in Armenia, some of which are already under way, need to mature. Central to the policy task are improvements in what may be called the innovation system (that is, the network of organizations, rules, and procedures that affect how a country acquires, creates, dis- seminates, and uses knowledge). Traditional measurements for an innovation system include indicators of expenditure on R&D, activ- ity in high-technology sectors (biotechnology, ICT), patenting activity (numbers, intensity), and number of researchers per 10,000 population. These indicators proxy the ability to generate new knowledge. Key organizations for the creation of knowledge include universities, pub- lic and private research centers, and policy think tanks. Private firms are at the center of the innovation system. If the private sector has little demand for knowledge, the innovation system cannot be effec- tive. Effective R&D-industry linkages are vital to transform knowl- edge into wealth. Therefore, networking and interactions among the different organizations, firms, and individuals are critically important. The intensity of these networks, as well as the incentives for acquir- ing, creating, and sharing knowledge, are influenced by the economic incentive regime in general. Acronyms and Abbreviations ACF Average Cost of Funds ADA Armenia Development Agency AIA Armenian International Airlines ASA Air Service Agreements ASYCUDA Automated System for Customs Data ATC Air Traffic Control BA British Airways BEEPS Business Environment and Enterprise Performance Survey CAA Civil Aviation Administration CEEC Central and Eastern European Countries CEM Country Economic Memorandum CES Constant Elasticity of Substitution CGE Computable General Equilibrium CIF Cost, Insurance, and Freight CIS Commonwealth of Independent States CMEA Council for Mutual Economic Assistance CNC Computer Numerical Control CPEC Commission for the Protection of Economic Competition CPI Corruption Perception Index CRS Computer Reservation Services DSA Debt Sustainability Analysis EASA European Aviation Safety Agency EBRD European Bank for Reconstruction and Development EC European Commission ECA Europe and Central Asia EDC Electricity Distribution Company EIF Enterprise Incubator Foundation EIR Economic and Institutional Regime/Economic Incentive Regime xxxi xxxii Acronyms and Abbreviations EQ&R Education Quality and Relevance ERC Energy Regulatory Commission EU European Union FAT Foreign Affiliate Transfers FDI Foreign Direct Investment FIAS Foreign Investment Advisory Service FSU Former Soviet Union GAMS General Algebraic Modeling System GATS General Agreement on Trade in Services GATT General Agreement on Tariffs and Trade GDCA Director General of Civil Aviation GDP Gross Domestic Product GDS Gross Domestic Savings GERD General Expenditures on Research and Development GNI Gross National Income GNS Gross National Savings GSM Global System for Mobile Communications GTAP Global Trade Analysis Project GVA Gross Value Added HDI Human Development Index HH Household HSBC Hong Kong and Shanghai Banking Corporation IAP International Aviation Policy ICAO International Civil Aviation Organization ICT Information Communication Technology IDA International Development Association IEF International Enterprise Foundation IFC International Finance Corporation IFRP Integrated Financial Rehabilitation Plan ILCS Integrated Living Conditions Survey ILO International Labour Organization IMF International Monetary Fund IO Input-Output IP Internet Protocol IPR Intellectual Property Rights ISAM Internal Single Aviation Market ISIC International Standard Industrial Classification ISP Internet Service Provider ITA International Technology Agreement JAA Joint Aviation Authorities JAC Junta de Aeronáutica Civil de Chile KAM Knowledge Assessment Methodology KE Knowledge Economy KEI Knowledge Economy Index K4D Knowledge for Development KI Knowledge Index LCC Low-Cost Carriers Acronyms and Abbreviations xxxiii LHS Left-Hand Side LIL Learning and Innovation Loan MCA Millennium Challenge Account ME Maximum Entropy MFN Most Favored Nation MoU Memorandum of Understanding MTEF Medium-Term Expenditure Framework NBER National Bureau for Economic Research NSS National Statistical Service OECD Organisation for Economic Co-operation and Development PPORF Practical Program of Revolution in Factories and Other Organizations PPP Purchasing Power Parity PRSP Poverty Reduction Strategy Paper PSRC Public Services Regulatory Commission R&D Research and Development RA Representative Agent RCA Revealed Comparative Advantage RCS Regulatory Cost Survey REER Real Effective Exchange Rate RHS Right-Hand Side S&T Science and Technology SAC Structural Adjustment Credit SAM Social Accounting Matrix SAS Scandinavian Airline Systems SEE Southeastern Europe SIMA Statistical Information Management & Analysis SITC Standard International Trade Classification SME Small and Medium Enterprises SNA System of National Accounts SOE State-Owned Enterprise SRI Structural Reform Index SWOT Strengths, Weaknesses, Opportunities, Threats TCS Teaching Company Scheme TFP Total Factor Productivity ULC Unit Labor Cost UNDP United Nations Development Programme USAID United States Agency for International Development USPTO United States Patent and Trademark Office VAT Value Added Tax VoIP Voice over Internet Protocol WBI World Bank Institute WDI World Development Indicators WITS World Integrated Trade Solution WTO World Trade Organization YBF Yerevan Brandy Factory Part I: Policies to Sustain Economic Growth Chapter 1 The Pattern of Growth A rmenia's macroeconomic performance has been a success.1 Mod- erate but robust economic growth in the initial years of the recov- ery (5 percent on average over 1994­2000) accelerated in the late 1990s, reaching an average annual rate of 11 percent since 2001. Growth has been fostered by a sound macroeconomic stance and a steady pur- suit of first-generation structural reforms, and has relied on excep- tionally high foreign assistance. Poverty began to fall at the end of the 1990s. Yet unemployment has persistently remained at double-digit levels and every third Armenian was still below the overall poverty line in 2004.2 On average, roughly one out of five economically active people is unemployed (based on survey data) and two-thirds of formal work- ers depend on incomes from low-productivity agriculture and trade. The brain drain had not been reversed as of 2005, indicating that there are limited employment opportunities across a wide range of skills. As in other transition economies, the formal labor market has seen strong gains in real wages based on rises in productivity but little increase in employment; informal labor markets have been stagnant. The impact of growth on poverty and extreme poverty has been strong. The initial phase of growth (1994­2000) had a moderate and uneven impact on poverty reduction, but, with the sharp accelera- tion in growth since 2000, poverty has fallen massively. The favora- ble effects from recent growth have been uneven across the regions, with the greatest benefit occurring in Yerevan. Urban areas outside of Yerevan, in particular, have experienced a lesser improvement. The present growth pattern rests on concentrated and volatile sources. The impact of growth on the generation of employment continues to be disappointing. 3 4 Part I: Policies to Sustain Economic Growth Growth has been dependent on foreign aid. During the entire trans- formation period Armenia has benefited from exceptionally generous international assistance (Figure 1.1) and has also been aided by a high level of remittances and private transfers from diaspora Armenians. Income and investment in Armenia continue to rely heavily on for- eign savings, while gross domestic savings, although on an increasing trend, are still low at about 9 percent of GDP. Financial deepening and the economic diversification of industry, manufacturing, and services have not taken off (see Figure 1.2 for a comparison with other Europe and Central Asia [ECA] countries). Agriculture's contribution in Armenia to GDP is around 30 percent, Figure 1.1 Official Development Assistance, Cumulative 1992­2003 Albania Kyrgyz Republic Armenia Bolivia Tajikistan Georgia Moldova Azerbaijan Bulgaria Estonia Lithuania Latvia Paraguay Poland Uzbekistan country Ukraine Turkmenistan Romania Belarus Slovak Republic Kazakhstan Russian Federation Cyprus Czech Republic Hungary Thailand Croatia Slovenia Turkey 0 50 100 150 200 percent of GDP Source: World Development Indicators (WDI). The Pattern of Growth 5 Figure 1.2 ECA Countries, Income per Capita and Employment in Agriculture, 2003 60 Kyrgyz Rep. Georgia total) 50 Moldova of Armenia (% Romania 40 Azerbaijan Uzbekistan Turkey 30 agriculture Bulgaria in Kazakhstan 20 Poland Lithuania Ukraine Greece Latvia Croatia Portugal Russian Federation 10 Hungary Spain Ireland employment Estonia Norway Czech Rep. Cyprus Germany 0 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 11.0 GNI per capita, PPP (current international US$) Source: WDI. and to employment it is more than 45 percent, well above the levels of 4.5 percent and 20 percent, respectively, in new EU member states. The services sector, concentrated in public administration and trade, generates about 35 percent of GDP in Armenia compared to more than 60 percent in new EU member countries. As a small develop- ing economy Armenia depends critically on external markets. With a population estimated at 3.2 million and a GDP per head of about US$1,600 distributed quite unevenly, Armenia faces considerable chal- lenges in both development and poverty reduction. Armenia is a land-locked country with high transportation costs. Two important neighbors, Azerbaijan and Turkey, closed their bor- ders with Armenia for political reasons soon after Armenia's inde- pendence. Border closures have imposed costs on the economy, with the main surface trade link limited to (i) low-capacity rail and road connections with Georgia and its Black Sea ports, and (ii) Iran via a single road. High transport costs arise from border closures, but also, importantly, from policy weaknesses in the transport and communi- cations regimes. Overcoming policy-induced and structural problems such as these makes development and poverty reduction challenges even more complicated. 6 Part I: Policies to Sustain Economic Growth SUSTAINED GROWTH WITH IMPRESSIVE PRODUCTIVITY GAINS As noted, Armenia's macroeconomic performance during the transition has been very successful. A sharp 60 percent decline in output between 1991 and 1993 was reversed by a rapid recovery. Since 1994, Armenia has grown at a remarkable average annual rate of over 8.0 percent, and by 2005 its real GDP reached 120 percent of its pretransition level. This growth pattern is comparable with the Central and Eastern European (CEEC) and Baltic countries, where the resumption of growth dates from 1996­97,3 rather than with the CIS (Figure 1.3). Growth was robust for the entire past decade (Figure 1.4). It started moderately in the first years of economic recovery (1994) and accelerated toward the end of the decade in response to the strong contribution from investments. The macroeconomic impact of the regional financial crisis (1998­99) was moderate. The authorities have mitigated the negative impact by the timely implementation of pre- cautionary measures. In addition, the initiation of diamond exports and subdued imports have compensated for the temporary erosion of other sources of growth in 1999. The regional crisis, however, had a strong adverse impact on the terms of trade for agriculture, with an associated impact on pov- erty. It also dampened remittances with a painful effect on house- hold incomes. Economic growth accelerated in 2000 and was led by growth in consumption, while net exports contribution was small as both exports and imports showed strong expansion (Figure 1.5). As noted, growth has been dependent on foreign aid. During the entire transition period Armenia has benefited from exceptionally generous international assistance and has also been aided by a high level of remittances and private transfers from diaspora Armenians.4 The current account balance has seen considerable improvement as a result of steady reduction in trade deficit (Figure 1.6). The foreign trade and exchange regime was liberalized in the early 1990s and has contributed to promoting trade and investment. In addition to a pru- dent monetary policy, the large import content of the consumer price index also helped to hold down the rate of inflation when import prices fell in the late 1990s. Total factor productivity, reflecting efficiency gains from macroeco- nomic stabilization and structural changes, has driven GDP growth in Armenia. During the 1990s and early 2000s, Armenia realized much of the "catch up" potential which arose following the economic trans- formation. Labor productivity growth outperformed GDP growth and capital productivity growth was also high. As a result, the productiv- ity gap between Armenia and the industrialized countries has nar- rowed, but it remains substantial (Figure 1.7). The Pattern of Growth 7 Figure 1.3 Index of Real GDP, 1990­2005 250 Armenia 200 CEEC and Baltics 150 1994=100 100 ECA region CIS 50 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004p2005p year Source: World Bank, ECA Regional Tables. Figure 1.4 GDP Growth Rate, 1996­2006 35 gross capital formation 30 25 exports of goods and services 20 15 cent per 10 GDP 5 0 total consumption -5 imports of goods and services -10 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005p year Source: NSS. 8 Part I: Policies to Sustain Economic Growth Figure 1.5 GDP Growth and Contribution of Expenditure to GDP, 1996­2005 20 15 consumption 10 points GDP 5 investment 0 percentage -5 resource balance -10 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005p year Source: NSS. Figure 1.6 Evolution of REER, Trade Balance, and Current Account Balance, 1999­2005 120 0 real effective exchange rate 100 (left-hand side) -2 80 -4 -6 1997=100 60 -8 GDP 40 current account balance of index, (right-hand side) -10 20 -12 percent 0 GDP -14 of -20 -16 percent -40 -18 1999 2000 2001 2002 2003 2004 2005p trade balance (left-hand side) Source: NSS. CAB = current account balance; REER = real effective exchange rate. Figure 1.7 Relative Productivity Level and Productivity Growth Rate, Latest Data 45 8 40 6 2004=1 35 in 30 4 level 25 2 20 percent productivity 15 0 10 -2 Armenia 5 The 0 -4 Pattern de China Korea Greece ChinaChile France JapanItaly Peru R.B. ArmeniaEstonia Latvia GeorgiaSloveniaof Finland Ukraine SwedenNorwayPortugalBulgariaPakistan Turkey BoliviaSpain Mexico Moldova Romania Ecuador Azerbaijan KazakhstanCroatia PolandHungarySingapore Federation Australia AustriaDenmarkKingdom GermanyColombiaBelgium IndonesiaSalvadorUruguay Argentina Kong, Netherlands El of Republic United Russian Hong Venezuela, Gr owth relative productivity level (left-hand side) productivity growth rate (right-hand side) Source: WDI, Staff Calculations. 9 Note: Data are for 2004; 2003 data are used if 2004 data are not available. 10 Part I: Policies to Sustain Economic Growth Armenia's unit labor cost (ULC) generally compares favorably to neighboring markets. This is mainly due to Armenia's comparative advantage in having an educated and skilled workforce and the coun- try's strong tradition of highly skilled craftsmen. In the mid-1990s the real wage increase outperformed the increase in labor productivity. This was partially due to the wage adjustment from a very low base. Unit labor costs did not rise during the 1990s as the real exchange rate depreciated. The early 2000s were characterized by strong productiv- ity gains and, except in 2003, outperformed the increase in real wages. The unit labor cost started to increase in late 2002, reflecting both the increase in employment compensation and the real effective exchange rate (REER) appreciation and declined again in 2005 in response to wage stability and strong productivity growth. Over the 1990s, the unit labor cost increased to about 46 percent of the U.S. level (about 44 percent of the euro-area level) from as low as less than 3 percent in the early 1990s (Figure 1.8). Figure 1.8 Relative Unit Labor Cost Index, 1997­2004 106 48 105 46 104 103 44 102 42 101 States=100 40 100 percent 99 38 United 98 36 97 34 96 95 32 1997 1998 1999 2000 2001 2002 2003 2004 year United States (left-hand side) Euro area vs U.S. (left-hand side) Armenia vs U.S. (right-hand side) Armenia vs Euro area (right-hand side) Source: OECD, WDI. The Pattern of Growth 11 Productivity gains have risen because of the enterprise restructur- ing and the contemporaneous macroeconomic discipline that ena- bled enterprises, through competition and trade, to realize efficiency improvements in the areas of management, marketing, and technol- ogy. They have also risen because of labor shedding from declining economic activities and sectors and the absorption of labor by new enterprises. In contrast, productivity fell in agriculture for reasons dis- cussed earlier. Armenia's exchange-rate regime has been stable since the national currency was introduced, and the REER depreciated gradually until 2004 (Figure 1.9). The strong appreciation of the dram (the Arme- nian currency) since early 2004, in response to sustained high rates of growth and productivity gains in the economy (the Balassa-Samuelson effect), has ensured the continued importation of price disinflation to the benefit of economic performance. INCREASINGLY EXPORT-LED GROWTH Following an initial prolonged period of export weakness, partly owing to adverse geopolitical conditions peculiar to Armenia, extremely strong export growth has been registered since the late 1990s. This development reflects a transition to market conditions in three ways. First, exports have been based on restructured industrial capacities. The inherited industrial capacities that were incompatible with market disciplines have been dismantled. No other CIS econ- omy recorded such a strong growth in 1999­2004. Armenian firms have also outperformed other non-oil CIS competitors in both CIS and EU markets, with their shares in respective markets increasing very significantly in the 1999­2004 period. The export share in total EU imports doubled in 2003 alone. Against the background of falling Russian imports from most CIS countries, the Armenian performance is particularly impressive, with the value of exports of goods more than doubling between 1999 and 2004. Trade with the CIS appears no longer driven by the post-Soviet hysteresis in trade patterns, with the emerging trade reflecting a comparative advantage in these increas- ingly competitive and market-oriented markets. Second, the readjustment in the geographical pattern of trade, reflecting the economic weight of regional markets, appears to be com- plete. While in 1995 Russia, together with other CIS countries, took 56 percent of Armenia's exports and supplied 49 percent of Armenia's imports of goods, these shares fell to 25 percent and 27 percent, respec- tively, in 1999 and to 19 percent and 29 percent in 2005. The shift has been largely toward the EU-15, whose share in Armenian exports rose 12 Part I: Policies to Sustain Economic Growth Figure 1.9 REER Quarterly Increase, Indicating Appreciation 115 110 105 100 95 appreciation 90 85 80 Q1-97 Q1-98 Q1-99 Q1-00 Q1-01 Q1-02 Q1-03 Q1-04 Q1-05 quarter Source: Central Bank of Armenia. from 26 percent in 1995 to 46 percent in 2005 and in imports from 15 percent to 26 percent over the same period. Third, the expansion in exports has not been confined to goods but has also included services, especially if the estimate that puts exports of ITC services at around US$100 million rather than the US$11 mil- lion reported in the balance of payments category of "computer and information services" is broadly correct. With or without a revised figure for these services, revenues from services increased more than expenditures. With a revised figure, the balance of trade in services swung to the surplus in 2003, and the overall deficit in trade in goods and services was significantly lower. Finally, Armenia's growth has become export-led, as exports growth has far outstripped GDP growth. (Figure 1.10). Moreover, a strong GDP growth performance considerably lessened an overall potentially adverse macroeconomic impact of trade imbalances. In terms of the GDP, the trade deficit fallen rather rapidly from 38 percent in 1997 to 12 percent in 2005. The Pattern of Growth 13 Figure 1.10 Key Economic Indicators of External Performance, 1998­2005 60 50 40 30 cent per 20 10 0 -10 1998 1999 2000 2001 2002 2003 2004 2005 year exports as percent of GDP annual rate of GDP growth imports as percent of GDP annual rate of export growth Source: NSS. The composition of exports of goods has registered some encour- aging developments, with exports diversifying toward high value­ added goods. Yet a significant weakness is that manufactures, other than diamonds, have been conspicuously absent in Armenia's exports. Although the share of manufactures increased in Armenia's total exports from 39 percent in 2000 to close to 70 percent in 2005, this was mainly due to the increase in exports of diamonds. Other manu- factured exports fell in terms of value with their share in total exports falling from 29 percent in 2000 to 10 percent in 2003 alone. CONTINUING HEAVY RELIANCE ON EXTERNAL SAVINGS Armenia's heavy reliance on foreign savings to finance its investment needs has declined over time but remains large (Figure 1.11). In 2005 about 20 percent of total investment was financed through foreign 14 Part I: Policies to Sustain Economic Growth Figure 1.11 Savings and Investment Balance, 1994­2005 30 total investment 20 net current transfers 10 foreign savings GDP GDS of 0 cent per -10 net income from abroad -20 -30 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 year Source: NSS. GDS = gross domestic savings. savings as compared to 90 percent in 1999. This reduction in reli- ance on foreign savings did not entail a decline in investments, as a commensurate rise in gross domestic savings has taken place. But the investment share in GDP has risen only modestly (growth has been driven by productivity) and the rise in investment necessary for sustained growth will require a greater domestic savings effort. Total investments averaged 19 percent of GDP during 1996­2000, increasing to 23 percent of GDP in 2001­05, but Bank projections show that the investment-to-GDP ratio will need to rise to around 28 percent by the end of the decade to sustain growth, assuming a return to a steady rate of productivity growth. Gross domestic savings remain low as shown by international com- parisons; and gross national savings fall short of those in some other CIS countries. Though the external debt and current account posi- tions are comfortable, these figures illustrate the dependence of the economy on income and transfers from abroad to finance investment needs (Figure 1.12). The Pattern of Growth 15 Figure 1.12 Gross Domestic Savings and Gross National Savings, 2004 Albania Armenia Georgia Bulgaria Kyrgyz Republic GNS Poland GDS Romania Lithuania Latvia Hungary country Croatia Ukraine Slovak Republic Estonia Slovenia Czech Republic Russian Federation Azerbaijan Kazakhstan 0 5 10 15 20 25 30 35 40 percent of GDP Source: World Bank ECA Regional Tables. GDS = gross domestic savings; GNS = gross national savings. Investment has been broad-based, but with a tilt toward housing construction. The structure of investments by sectors during 1999­2005 (Figure 1.13) shows that housing represents over 30 percent of total investment. Infrastructure has also attracted a considerable share of investment. Several profitable and competitive traditional sectors have attracted private investment during the process of economic transfor- mation (for example, gem cutting, brandy distillation, electric motors, hotels, high-technology manufacturing, and software services). A sig- nificant source of investment financing has been foreign direct invest- ment (FDI), but associated with large transactions. A large part of the 16 Part I: Policies to Sustain Economic Growth Figure 1.13 Sectoral Shares in GDP and Investment by Sectors, Average, 1999­2005 35 30 25 20 percent 15 10 5 0 industry agriculture transport trade housing other sector share in investment 1999­2003 share in GDP 1999­2005 Source: NSS. FDI, particularly in infrastructure (telecommunications through the privatization of the national telecommunications company, ArmenTel) and in industry (privatization of the gas distribution network) was the direct result of large-scale privatization. COMFORTABLE OUTLOOK FOR EXTERNAL DEBT A debt-sustainability analysis (DSA) carried out for Armenia shows that the country is considered to be at a low risk for debt distress, with all debt indicators well below the relevant country-specific debt-burden thresholds, including when subjected to stress tests (see Table 1.1). The baseline DSA has been developed on a cautious set of assump- tions, including: (i) a real GDP growth of 5 percent per year over the period 2008­12 and 4 percent thereafter; (ii) an inflation rate of 3 percent throughout the projection period; (iii) an overall fiscal deficit of 2.9 percent of GDP through the Poverty Reduction Strategy Paper The Pattern of Growth 17 Table 1.1 Armenia's External Debt Burden Indicators Armenia's Ratios Thresholda 2005 2025 NPV of debt in percent of: Exports 200 60 25 GDP 50 15 8 Debt service in percent of: Exports 25 6 2 Source: IMF-WB Staff calculations a. Threshold values are based on the guidelines for low-income country DSAs, among which Armenia is considered to be a strong performer. NPV = net present value. (PRSP) horizon of 2015 and 2.3 percent of GDP thereafter (implying a debt-stabilizing primary balance); (iv) a steady increase in the share of both exports and imports as a proportion of GDP, leading to a gradual improvement in the current account balance over the projec- tion period; and (v) a shift away from highly concessional forms of external financing. Under the baseline scenario the fiscal position remains sustainable. The net present value (NPV) of the ratio of debt to revenue and NPV of the ratio of debt service to revenue remain relatively stable through- out the projection period. In particular, the NPV of external debt falls steadily from around 15 percent of GDP in 2005 to just over 8 per- cent by 2025, the NPV of the debt-to-exports ratio falls from just over 60 percent to around 25 percent over the same period, and the debt service-to-exports ratio declines from just under 6 percent in 2005 to around 2 percent by the end of the projection period (Figure 1.14). The framework for low-income country DSAs also incorporated alternative scenarios and bound tests aimed at identifying the sensi- tivities of the baseline projection to a range of potential shocks. Based on these scenarios and tests it can be seen that Armenia's debt outlook is particularly sensitive to (i) a lasting shock to the growth outlook and (ii) a slowdown in export growth. Nevertheless, even under these scenarios, Armenia's debt outlook is likely to remain manageable. The DSA also considered the possible impact on debt sustainability of a "significant" fiscal event, such as the assumption by the govern- ment of pension liabilities. Assuming such an event was to occur in 2010, costing around 10 percent of GDP, the NPV of the public debt- to-GDP ratio in 2025 would remain manageable at about 40 percent. While the historical scenario (which can be considered as a "no-reform" 18 Part I: Policies to Sustain Economic Growth Figure 1.14 Indicators of Public and Publicly Guaranteed External Debt under Alternative Scenarios, 2005­25 30 NPV of debt-to-GDP ratio 25 20 15 percent 10 5 0 2005 2007 2009 2011 2013 2015 2017 2019 2021 2023 2025 year 90 NPV of debt-to-exports ratio 80 70 60 50 40 percent 30 20 10 0 2005 2007 2009 2011 2013 2015 2017 2019 2021 2023 2025 year 7 debt service-to-exports ratio 6 5 4 percent 3 2 1 0 2005 2007 2009 2011 2013 2015 2017 2019 2021 2023 2025 year baseline historical scenario most extreme stress test Source: IMF-WB Debt Sustainability Analysis for Armenia. The Pattern of Growth 19 scenario) also projects debt ratios below the threshold indicators, the ratios generally continue on an upward trend through the projection period. This finding underlines the importance of continuing the proc- ess of structural reform and preserving macroeconomic stability in order to safeguard the debt outlook. KEY MACROECONOMIC CHALLENGES Despite the success of macroeconomic management thus far and the sustainability in medium-term fiscal and external accounts, the author- ities have to remain vigilant regarding the major sources of vulner- ability to economic performance. Economic projections to the year 2010 indicate that even with some moderation of growth to an average of 6 percent per year, external debt and debt service indicators will fall sharply, thereby creating both fiscal space and greater private sector funds for financing investment. On reasonable assumptions of produc- tivity growth and incremental capital-output ratios, the investment-to- GDP ratio will have to rise by around 4 percentage points in the rest of this decade. This rise in the investment-to-GDP ratio is projected to be financed equally by external savings (as grant inflows into the economy rise substantially) and private sector savings. Thus, the macroeconomic management challenge will lie in ensur- ing stability that leads to the generation of greater private savings and efficient use of official inflows. The task of monetary and exchange rate policies will be to bear down on inflation, which has flared up periodically in recent years. The appreciation of the dram that is under way will help to fight inflation. In an economy with rapid productivity growth such as Armenia's, a real appreciation of the currency is to be expected and is the mechanism for establishing a balance between the tradables and the nontradables sectors. The Central Bank is entirely right in permitting the nominal appreciation seen thus far: were it to seek to hold the nominal rate down, the real appreciation would be realized through higher inflation. It is particularly important to permit the exchange rate to be determined entirely through market forces, as the bank lacks the long-duration financial instruments necessary for sterilization activities. One-year bills currently being issued will help in short-term liquidity control and in operations to dampen day-to- day volatility in the exchange rate. The fiscal deficit over the medium term should be confined to a range of 2.5 to 3 percent of GDP. This will require some broadening of the fiscal base and a rise in the revenue mobilization ratio to finance priority social expenditure needs. The narrowness of the fiscal base and the insufficient reliance on direct taxes lead to a lack of balance in the tax structure and to a mobilization ratio that is too low to fund 20 Part I: Policies to Sustain Economic Growth priority social expenditure needs. Changes in tax policy to enhance the role of direct taxes5 and in tax administration to ensure minimal evasion and fair treatment of all taxpayers will go a long way to sta- bilizing the fiscal base. As labor taxes and social contributions are high, particularly given the high unemployment and the large infor- mal economy, consideration could be given to a shift toward direct taxes. Any temptation to relax on tax reforms against a background of large donor grant inflows should be resisted. Questions also arise as to the capacity of the economy to efficiently absorb the large volume of the external grant funds flowing from the Millennium Challenge Account (MCA) and other sources. Constraints to efficient absorption arise from macroeconomic factors, institutional and managerial factors, and infrastructure. Projected inflows equiva- lent to 2 percent of GDP will not be inconsistent with the macro- stability provided that the monetary and exchange rate stance is as discussed above, the import content of the counterpart to the grant financing is high (as would be the case with infrastructure capital spending), and domestic spending (for example, on salaries) is con- trolled. In this manner, the inflationary impact and the real apprecia- tion of the dram will be minimized. Macroeconomic challenges will also arise if grant financing is substituted for domestic revenue mobi- lization or if it leaks into public consumption--these situations would pose a danger to fiscal sustainability over the medium term. On the institutional side, the appropriate budgeting of inflows and the selection of projects to reflect high-priority needs as identified in the PRSP action plan, as well as careful management of the public investment program, are essential if absorption is to be efficient. More- over, infrastructure constraints to efficient absorption will be eased if priority is given--as envisaged--to addressing needs in the transport and water sectors (especially in rural areas). There is a concomitant need to ensure high quality in public expen- ditures. Over 2005­06, a large increase in teacher salaries took place as salaries were grossly inadequate to attract quality services. This increase is related to productivity increases (the pupil-teacher ratio rises). In subsequent years, expenditures must shift from wages to items that enhance service quality as envisaged in the medium-term expenditure framework, such as curricular reform and education aids, the broadening of education services, primary health care, and increasing the real rate of the poverty benefit as well as targeting the benefit more sharply. A strategic approach to capital spending based on an overall public investment plan that explicitly makes sec- tor choices is also necessary, particularly in the context of the MCA disbursements. The Pattern of Growth 21 GROWTH AND A PRONOUNCED FALL IN POVERTY The most recent period of growth has resulted in a sharp, dispropor- tionate rise in consumption by the poor (Table 1.2).6 Growth incidence curves are useful for analyzing the impact of aggregate economic growth across households, as they plot the growth rate in consump- tion for individuals ranked according to their consumption. Figure 1.15a­d presents growth incidence curves for the period from 1998­99 to 2004. The vertical axis is ordered by increasing the levels of per- adult equivalent consumption. A downward sloping incidence curve indicates that people in the poorest quintiles of the population have benefited from growth more than the average. The horizontal line parallel to the horizontal axis indicates the growth rate in mean. In 2004 per adult equivalent consumption was about 10 percent higher than consumption in 2001, while output in Armenia grew by a cumulative 30 percent during the same period. This trend was not uniform across the capital and noncapital urban and rural areas. Over- all, this pattern appears to be driven by the distribution of the growth impact in the capital city and the rural areas. In contrast, with the exception of the people below the tenth percentile, whose consump- tion increased by 10 percentage points on average, the consumption of the non-Yerevan urban population was either unchanged or declined in 2004 compared to 1998­99. The population across all deciles, with the exception of non-Yerevan urban areas, experienced consumption growth, except for the richest population above the ninetieth percentile, which experienced a nega- tive growth rate everywhere except in the capital city. The poorer quin- tiles of the income distribution gained more from economic growth than the richer quintiles. The extremely poor have gained the most Table 1.2 Armenia's Poverty Indicators, 1999­2004 (percent) Poverty incidence (%) All Urban Yerevan Other urban Rural Extreme poverty 1998/99 26.1 32.1 29.6 34.5 18.0 2004 6.4 7.5 6.1 9.2 4.4 Overall poverty 1998/99 56.3 62.7 58.7 66.5 47.7 2004 34.6 36.4 29.2 43.9 31.7 Source: Integrated Living Conditions Survey (ILCS); Armenia 1998/99­2004. 22 Part I: Policies to Sustain Economic Growth Figure 1.15 Growth Incidence Curves, 1998/99­2004 a. Armenia: Growth Incidence Curve, 1998/99-2004 50 % growth incidence curve 40 adult 04- per 30 98/99 mean of growth rates cumulative rate 20 growth rate in mean equiv. 10 consumption growth 0 0 20 40 60 80 100 consumption per adult equivalent percentiles b. Armenia: Growth Incidence Curve in Yerevan, 1998/99-2004 50 growth incidence curve % 40 mean of growth rates adult 04- per growth rate in mean 30 98/99 cumulative rate 20 equiv. 10 consumption growth 0 0 20 40 60 80 100 consumption per adult equivalent percentiles from the recent economic growth. This evidence is consistent with the evolution of inequality for the period--the Gini coefficient decreased from a value of 0.32 in 1998­99 to a value of 0.23 in 2004--and indi- cates that the growth was accompanied by a compression in the over- all structure of the consumption distribution. This favorable pattern has been driven by the increase in nonfactor income and indicates the effectiveness of the social protection policy through the improved The Pattern of Growth 23 Figure 1.15 (continued) c. Armenia: Growth Incidence Curve in Other Urban Areas, 1998/99-2004 50 % growth incidence curve 40 adult 04- per mean of growth rates 30 98/99 cumulative growth rate in mean rate 20 equiv. 10 consumption growth 0 0 20 40 60 80 100 consumption per adult equivalent percentiles d. Armenia: Growth Incidence Curve in Rural Areas, 1998/99-2004 60 % growth incidence curve 40 adult 04- per 20 mean of growth rates 98/99 cumulative rate 0 growth rate in mean equiv. consumption -20 growth -40 0 20 40 60 80 100 consumption per adult equivalent percentiles Source: ILCS; Armenia 1998/99­2004. Note: The curve refers to the period of 5 years and 9 months. targeting of the family poverty benefit system and the increased social transfers (such as pensions, family poverty benefits, and others), and also points to the effect of remittances and the multiplier effect from externally driven demand sources. The consumption figures shown above fit in well with the data on poverty indicators. The strong gains for both poverty and extreme pov- erty in Yerevan are consistent with growth rates of about 17 percent 24 Part I: Policies to Sustain Economic Growth Table 1.3 Armenia: Growth and Inequality Decomposition of Changes in Poverty Incidence between 1998­99 and 2004 (average effects) Total Urban Yerevan Other urban Rural Extreme poverty Change in poverty incidence ­20.6 ­25.9 ­26.3 ­25.4 ­13.5 Growth component ­14.2 ­22.6 ­21.4 ­21.7 ­4.1 Redistribution component ­6.4 ­3.4 ­4.8 ­3.6 ­9.5 Overall poverty Change in poverty incidence ­24.3 ­32.0 ­36.7 ­26.8 ­13.8 Growth component ­22.5 ­31.1 ­33.3 ­27.4 ­7.8 Redistribution component ­1.8 ­0.9 ­3.4 0.6 ­6.0 Source: ILCS; Armenia (various years). in the mean consumption of the bottom three deciles, as compared to about 6 percent for the same deciles in the non-Yerevan urban areas and 12 percent in the rural areas. Armenia's poverty is skewed toward being an urban problem (outside of the capital). The reduction in poverty is mostly driven by consumption growth. The decomposition of poverty reduction into a growth and a distribu- tional component, developed by Datt and Ravallion (1992), gives bet- ter insight into the poverty effects of decreasing inequality (Table 1.3). Calculations for both extreme and overall poverty incidence reveal that the growth and redistribution components worked in the same directions in influencing the reduction in poverty. However, most of the observed poverty decline is attributed to the growth component (around two-thirds, on average). Changes in consumption distribution played a more important role in extreme poverty reduction than in overall poverty reduction. Decreased inequality contributed about 31 percent to extreme poverty reduction and 7 percent to overall pov- erty reduction. Thus, the importance of the inequality component in explaining changes in overall poverty in Armenia is similar to that in other CIS countries where changes in inequality explained 5 percent of the changes in poverty (Alam et al. 2005). In the capital city of Yerevan and other urban areas, consumption growth was the most important source of the reduction in poverty (extreme and overall), as its contribution was over 80 percent. For rural poverty, on the other hand, both components--the decline in inequality and consumption growth--were almost equally important. Extremely poor rural households benefited mainly from the decline in inequality, as 70 percent of the reduction in rural poverty was attribut- able to decreased inequality. The Pattern of Growth 25 GROWTH WITHOUT JOBS An analysis of the impact of growth on employment is bedeviled by the change in economic regime. First, a market-based economy has replaced the command economy directed at realizing economies of scale on a narrow range of products--such as electronics and syn- thetic rubber--for dispatch to the rest of the Soviet Union. Second, upon independence the economy experienced large external price, output, and infrastructure shocks. In the initial phase (1992­94), there occurred a massive shift of labor to subsistence agriculture in response to the initial price and output shocks, and labor hoarding took place in enterprises that continued to function at considerably shrunken levels of output. The strong growth in output seen over the past decade has not resulted in a commensurate rise in employment until very recently. Output growth has been supported by strong labor productivity increases. Though registered unemployment stands at around 10 per- cent of the labor force, survey data point to rates twice as high, and the labor force has been affected by weak participation rates (Figure 1.16) and large migration, both skilled and unskilled. Much of the work force is to be found in subsistence agriculture or in the informal economy and underemployment is endemic. The weak demand for labor has persisted for a decade, and the rapid growth in productivity in formal activities has translated into rising real wages rather than more demand for labor. The labor market is segmented and has weak geographical mobility within the country. This phenomenon, called "jobless growth," is not specific to Arme- nia but is typical of most transition economies, where initial growth is mostly based on large productivity gains due to structural changes and labor shedding in the process of labor rationalization and more efficient utilization of labor resources rather than through an increased use of the labor force. According to a recent study, Armenia has actu- ally registered a significant "job loss growth" (Kuddo 2006). Over the last 15 years since 1990, employment has declined by 34 percent from 1.63 million people to around 1.1 million in 2004. In the process of the country's search for new equilibrium, the destruction of unproductive jobs has been accompanied by the creation of new and more produc- tive jobs. Job losses usually exceed job gains, leading to a net fall in employment. Only the World Bank's (2004a) Integrated Living Condi- tions Survey (ILCS) has recorded some increase in the employment rate among the population. Recent positive trends in employment indicate that Armenia may have reached a new equilibrium, and that growth will start to rely increasingly on improved labor utilization. The cities have been most deeply affected by transition, with labor participation and employment higher in rural areas. Other key 26 Part I: Policies to Sustain Economic Growth Figure 1.16 Participation, Employment, and Unemployment Rates in Urban and Rural Areas (population aged 15 and older) 80 participation rate 70 employment rate unemployment rate 60 50 40 percent 30 20 10 0 total Yerevan other urban rural sector Source: ILCS (2004) and NSS (2006). features of labor conditions in Armenia include the high percentage of youth that are neither in work or in study; the significant gender differences in participation; the extremely long duration of unem- ployment; the considerable regional disparities in participation; and the notable shift of employment from the state sector to the private sector in recent years. In comparison to the structural changes seen in the economies of the new members of the EU, those in Armenia have been much less profound thus far. The share of agriculture in GDP is today around 30 percent and agricultural employment accounts for more than 45 percent of total employment, well above the levels of 4.5 percent and 5 percent, respectively, in new EU member states. Activity is concen- trated in low value-added agriculture (dairy, meats, and grains). The services sector, concentrated in public administration and trade, gen- erates about 35 percent of GDP in Armenia, compared to more than 60 percent in new EU member countries. Subsistence agriculture, which is susceptible to supply shocks, remains overmanned, and those who rely on subsistence agriculture remain economically vulnerable. The Pattern of Growth 27 The impact of growth on formal job creation outside the construc- tion and trade sectors has been weak. Agriculture, which had been growing more slowly than the rest of the economy during the 1990s, has absorbed a surplus of (unskilled) labor released in the process of economic transformation. Part of the decline in the share of agriculture in total output from over 30 percent to 25 percent over the 1990s was related to changes in relative sector prices (that is, a decline in agricul- ture terms of trade since the 1998 Russian crisis) (Figure 1.17). Agricul- tural prices have been declining until recently, while prices in services (led by the utility price adjustment) and industrial prices have been consistently increasing. The special conditions that drove up the price of oil recently required another round of relative price adjustment. Agricultural prices saw another considerable decline in 2005. Clearly, such price shifts benefit the urban consumer and lead to the shift in resource allocation away from agriculture. But they have not led to the expected fall in employment in agriculture. In contrast, the fastest growing sector of the economy--construction--has made a significant contribution to job creation: the share of the labor force employed in construction has risen from 4 to 9 percent over the past five years. Job creation in industry in recent years was mostly concentrated in new private enterprises. Out of 26,500 jobs created in industry during 1999­ 2003, 22,000 were established in newly created private enterprises. Recent enterprise surveys indicate that there is a potential for employment growth. Average capacity use in Armenian firms, as Figure 1.17 Change in Sector Prices, Nominal Value-Added Deflator 1.7 construction 1.6 industry 1.5 GDP 1.4 services agriculture 1.3 1997=1 1.2 index,1.1 1.0 0.9 0.8 1997 1998 1999 2000 2001 2002 2003 2004 2005p year Source: NSS. 28 Part I: Policies to Sustain Economic Growth surveyed by the 2005 Business Environment and Enterprise Perform- ance Survey (BEEPS 2005), equals 82 percent, indicating that new jobs can be generated using the available capital stock. Moreover, employers point out that the optimal employment level in their firms compared to the current level implies an expansion of the workforce by 12.1 percent. This is more than the average in the ECA region of 8.6 percent of potential new employment in existing firms. GDP growth is expected to be more employment-intensive, since the rate of production growth in service sectors is expected to rise. Public employment has been constrained by the political choice of a small public sector in the economy. Other labor-intensive sectors and sectors with a demand for skilled labor, such as banking and light manufacturing, have been growing slowly since 2003 (Figure 1.18). The share of unskilled labor in total employment rose from 50 percent in 1990 to 75 percent in 2002. This development reflects some out-migration of skilled labor and possibly a worsening of the skills mismatch in the labor force after the collapse of central planning, as skills that were in demand under a planned economy were not needed under market conditions. (Labor-market research under way will examine such factors). Figure 1.18 Nonagricultural Jobs in the Productive State and Private Sectors, 1998­2003 500 jobs in productive state sector 450 jobs in productive private sector 400 350 300 250 thousands 200 150 100 50 0 2003 2002 2001 2000 1998 year Source: NSS and authors' calculations. The Pattern of Growth 29 Strong labor productivity growth and the increases in the ULC appear to have curtailed the potential for new employment in the short term. This trend has been reversed very recently with the increased reliance of growth on labor utilization. From an international perspec- tive, the Armenian labor market is still characterized by relatively low participation (65.9 percent compared to 70 percent in EU-19 and OECD) and employment rates, along with high unemployment rates (above 20 percent in 2004). In general, growth in labor productivity, if associated with slower growth in real wages, reduces ULCs and creates room for new employment. In Armenia, rapid growth in labor productivity was accompanied by increases in hourly compensation above the productivity growth level and related increases in ULC. The benefits of recent economic growth secured stronger earnings for current workers but did not generate higher employment, and the increase in labor costs contributed in part to jobless growth during the mid-1990s. Figure 1.19 illustrates the mismatch between the rapidly expand- ing sectors and labor productivity. The contrast between the rapidly growing sectors and formal employment indicates that the majority of the labor force has benefited only modestly from rapid economic expansion. The causes of the weak employment response to investment and growth lie in a business climate that discourages the flexible use of labor (business surveys provide evidence for this statement and are discussed in detail in Part 1, Chapter 3, of this book). Thus, a central policy task is to create conditions that will lead to an expansion of jobs. Raising the rate of job creation would require addressing the binding constraints to the formation and growth of firms. The prin- cipal weaknesses in Armenia are uncertainty in property rights, poor enforcement of contracts, arbitrary and predatory behavior on the part of public authorities, restrictiveness in the formal regime governing employment, and stunted financial intermediation. Considering that the labor market situation may be aggravated due to the fact that during the decade the labor supply will increase sig- nificantly, while the generation born at the birthrate peak in the 1980s will enter the labor force, Armenia will need to create more and better jobs. Thus, in addition to improving the business climate, Armenia will need to reform labor market institutions to create an adaptive labor market--that is, a market in which employers have the incentives to hire and workers have the incentives and required skills to take up available jobs. The upgrading of the labor force, or vertical mobility, is a precondition for rapid structural and technological change, for com- petitiveness, and for raising the share of high value-added products and services. The new role of employment services should be specified 30 Part I: Policies to Sustain Economic Growth Figure 1.19 Relative Labor Productivity by Sector 3.0 2.5 2.0 1.5 index 1.0 0.5 0.0 services industry average agriculture economic construction manufacturing sector 4.5 4.0 3.5 3.0 2.5 index 2.0 1.5 1.0 0.5 0.0 & stones power products banking economic average processing beverages mining quarrying & construction mineralprecious transportation food communication & & Source: NSS and authors' calculations. and the overall state of the labor market should be thoroughly ana- lyzed, including new forms of employment, underemployment, and unemployment; the components of labor force growth; age structure; industry and occupation structure; labor market segregation (for exam- ple, by age or gender); and regional imbalances. In contrast to some other CIS countries, the long record of stable macroeconomic management in Armenia reduces policy uncertainties. Moreover, with privatizations and enterprise restructuring being com- pleted, labor hoarding has practically been eliminated. In addition, wage-setting mechanisms are market-based and are decentralized to the firm level. However, employment protection legislation is restric- The Pattern of Growth 31 tive and dismissal costs are high in terms of severance pay, though enforcement is erratic (see Chapter 3 and Table 3.5). Thus, the muted employment response to growth reflects the incom- pleteness of structural reforms and of the development of institutions and practices that would encourage competition and enterprise activ- ity. To encourage employment growth, policy makers need to focus on property rights and all aspects of the rule of law, including rules- based regulation and regulatory behavior, bureaucratic arbitrariness and corruption, and the creation of a framework for bank and capital markets development. The persistent, very high rates of unemployment and the associated large, informal economy entail risks for the continuation of "tigerish" rates of growth. At a time when the memory of Soviet full employ- ment (even if of the "we pretend to work and they pretend to pay us" variety) and the adequate provision of very basic goods is still strong, the creation of haves and have-nots in the employment world--with large increases in real wages for a minority and with no clear prospect of a strong generalized growth in employment--diminishes support for the reform agenda still to be addressed. The strong association between unemployment and poverty makes reform even harder to promote. The worst affected by unemployment and poverty are middle- aged workers who may be less attuned by attitude and skills to the new economy. The remainder of this book is structured as an analysis of the policy response necessary to raise employment growth to the rates in com- parable tiger economies. Such an analysis will deal with questions of competition, the impediments to enterprise activity, the absorption of knowledge, the use of finance for investment, and the benefits of the international division of labor. SUMMARY Armenia's growth performance has been impressive and has had a sharp impact on poverty reduction. But growth has been depend- ent on foreign inflows (although the share of national savings has risen substantially), and growth has had a muted impact on lower- ing unemployment until very recently. The economy is insufficiently diversified; the export base remains narrow. The economy can be characterized as having a dual labor market: one formal, with rising productivity and real wages; the other, informal and stagnant. Thus, growth has been based on rising investment and productivity gains rather than on higher employment--a phenomenon seen in other CIS states. Broadening the base of growth and generating new jobs will 32 Part I: Policies to Sustain Economic Growth require changes in the business climate, sharper competition condi- tions, and a shift in taxation away from labor and social security taxes and charges and toward direct income taxation. NOTES 1. A background analysis for the pattern of growth can be found in Chap- ters 7 to 9 of Part II of this book. 2. Survey-based data show unemployment rates of around one-third; reg- istered unemployment is around 10 percent. In the EU-15 (the 15 EU mem- bers prior to the most recent enlargement of membership) and EU-10 (the 10 most recent members) less than 0.5 percent and 5.4 percent of the population, respectively, live below the poverty threshold of US$2 a day. 3. For a more detailed cross-country comparison see Loukoianova and Unigovskaia (2004). 4. Armenia has a very large diaspora of about 5 million compared to the country's population of about 3.2 million people. The diaspora is spread over the world and is large in the Middle East, the United States, Russia, and Western Europe. 5. Indirect tax rates are high (VAT [value-added tax] is levied at 20 percent) and the tax base is commendably broad; thus the scope for further action on indirect taxes is small. The tax system is discussed in Part 2 of this book. 6. A detailed study of developments in poverty can be found in ILCS; Armenia (various years). Chapter 2 Constraints to Sustained Growth THE RECORD As has been noted, Armenia has an impressive decade-long record of double-digit rates of growth on average, although incomes remain well below those in countries such as Lithuania. Armenia's strong growth record has been fostered by sound macroeconomic per- formance: namely, modest fiscal deficits, low inflation, and a stable (recently appreciating) currency. Exports have accelerated and the cur- rent account deficit has declined from over 20 percent of GDP in the mid-1990s to less than 5 percent in 2005. Gross international reserves remained at about four months of imports. Armenia has pursued a prudent debt management strategy. By now, almost all nonconces- sional debt, accumulated during the first years of independence, has been settled1 and the debt burden has been kept under control. Macroeconomic projections, on the assumption of continued stabil- ity-oriented policies, indicate a comfortable fiscal sustainability and plunging indicators for debt service and debt burden (as discussed in Chapter 1 of this book). Armenia has already implemented an impressive range of market- oriented reforms, including free price formation in a highly open market with a liberal regime for trade and investment, a liberal finan- cial system, total private ownership of land, and privatization of both small and medium enterprises (SMEs) and large enterprises.Acces- sion to the WTO in December 2002 has locked Armenia into a liberal foreign trade regime. The early and swift privatization of land was particularly impressive and stands in contrast to all other CIS coun- tries. Land ownership reforms have continued with an acceleration of land registration processes and streamlined titles issuance to farmers. These actions have been followed by the development of rural land 33 34 Part I: Policies to Sustain Economic Growth markets. The government adopted a privatization program in 2000 that resulted in privatization of 75 percent of the large public enter- prises and almost all of the SMEs. For those enterprises that were not sold following three attempts at privatization, the government established appropriate exit strategies for public enterprise reorgani- zation or liquidation. The government implemented several structural reforms in the energy sector, which included undertaking the financial and technical restructuring of the sector, increasing tariffs, establishing a regulatory framework, improving the collection of electricity bills, and reduc- ing cross-subsidization. The price of electricity reflects cost recovery. In addition, the government mitigated the adverse impact of tariff increases on the poor through increasing the well-targeted family benefits. Furthermore, the government privatized the Electricity Dis- tribution Company (EDC) in 2002. Since its privatization, the EDC has improved its financial performance significantly. The government has ensured that the privatized electricity distribution company will comply with its license. The government also carried out a wide range of reforms to improve the regulatory framework for utilities and infrastructure, especially for the power sector. The role and the independence of the Energy Regulatory Commission (ERC) were strengthened and a strategy was developed to improve the efficiency of the nonprivatized parts of the power sector. In addition, the government established a single utility regulator based on the ERC; adopted the necessary legal framework to ensure its financial autonomy; and transferred the economic regula- tions of the energy, drinking water, and telecommunications sectors to the single Public Services Regulatory Commission (PSRC). The gov- ernment also carried out performance monitoring and a public infor- mation program to ensure that people are made aware of the power sector's performance in a timely and credible manner. Other utilities also underwent deep reforms and experienced a gradual increase in private sector participation. The government adopted the Integrated Financial Rehabilitation Plan (IFRP) for pub- lic utilities--covering electricity, drinking water, irrigation, and urban electric transport--to guide the policy reforms and to ensure the ultimate commercialization of the utilities. The government made adequate provisions in the annual budgets to support the financial recovery plan of the transport and water sectors in order to ensure that full payments are made to the energy sector. Reforms in the water sector included the financial restructuring of drinking water compa- nies, tariff increases, the adoption of a new water code, and improved collections. In addition, the government adopted a schedule for irriga- tion tariff increases in 2002­07 to provide full cost recovery tariffs for the irrigation system by 2007. Constraints to Sustained Growth 35 The government undertook several reform measures in the banking sector to improve the lending environment. The government restruc- tured two weak and failing banks, adopted the law on the bankruptcy of banks, increased the minimum capital requirements of the banks, and modernized the loan classification system. In the area of banking supervision, the Central Bank developed manuals both on- and offsite, started on-site supervision, and introduced regulations for the inter- nal control of banks. The payments system was improved, diagnostic studies were completed that resulted in the revoking of the licenses of six banks, a credit registry at the Central Bank was introduced, and confidence in the banking system was enhanced by the introduction of a deposit guarantee system and a series of accounting standards. The government also adopted the law on insurance, took initial steps to strengthen creditor rights and corporate governance for banks, estab- lished a regulatory and supervisory environment for insurance, and adopted regulations and procedures for the Securities Commission. These reforms contributed to enhancing confidence in the banking system and accelerating credit growth. THE RECORD IN A COMPARATIVE CONTEXT Some interesting comparative insights can be gleaned from the exami- nation of European Bank for Reconstruction and Development (EBRD) transition indicators of progress (normalized for the purpose of this analysis) in the liberalization of prices, small-scale privatization, and foreign trade and exchange rate regimes (EBRD, various years).2 They all comprise measures at the core of the first-generation reforms (see Figure 2.1). Armenia and Lithuania began their journey toward a market-based economy from the same level, estimated at around 35 percent of the "desired" level. Nevertheless, while Lithuania made a big leap in 1993­94, Armenia initially lagged behind. However, both countries had largely completed first-generation reforms by 1998. It is interesting to note that the SEE-4 economies, although at the same starting point in terms of EBRD scores, have lagged behind Armenia in implementing first-generation reforms. CONSTRAINTS TO FUTURE GROWTH The preceding section of this chapter has argued that the factors under- pinning Armenia's impressive growth performance are durable stabil- ity in macroeconomic stance and the pursuit of first-generation struc- tural reforms, price and trade liberalization, a liberal foreign exchange regime, the privatization of all small enterprises and some medium Box 2.1 Armenia--Production Statistics by Sector Output Value added Sector Sector share share Sector US$ in total US$ in total Wheat, potatoes, and legumes 155.5 3.6 98.7 4.6 Vegetables and fruits including grapes and dried fruits 132.5 3.1 115.4 5.4 Vegetable oils and fats 15.8 0.4 1.0 0.0 Crops not elsewhere classified 103.1 2.4 81.1 3.8 Diary products including eggs and milk 342.6 7.9 168.2 7.8 Beef, pork, mutton, and poultry meat 159.0 3.7 66.9 3.1 Energy--oil and natural gas 75.7 1.7 0.0 0.0 Mining and quarrying 65.9 1.5 39.4 1.8 Food processing and beverages 430.7 9.9 158.1 7.4 Tobacco products 80.9 1.9 13.7 0.6 Light manufacturing and textiles 52.6 1.2 5.3 0.2 Chemicals, rubbers, and plastics 53.0 1.2 10.4 0.5 Mineral products and precious stones 269.3 6.2 40.7 1.9 Metals and metal products 109.9 2.5 43.6 2.0 Machinery, equipment, and motor vehicles and precision optical equipment 88.7 2.0 5.8 0.3 Other manufacturing 61.2 1.4 17.9 0.8 Electricity, gas, and water supply 152.0 3.5 111.6 5.2 Electricity supply and distribution 163.2 3.8 102.1 4.8 Construction 518.3 11.9 298.6 13.9 Transport and communications 259.9 6.0 153.1 7.1 Retail and wholesale trade and public catering 447.2 10.3 288.6 13.5 Banking, lending, and insurance 51.6 1.2 35.7 1.7 Governance, defense, and public procurements 362.5 8.4 197.7 9.2 Other services not elsewhere classified 93.7 2.2 23.6 1.1 Housing and dwellings 92.6 2.1 68.2 3.2 Total agriculture 908.6 20.9 531.3 24.8 Total manufacturing 1,036.4 23.9 251.9 11.7 Total industry 403.5 9.3 194.7 9.1 Total construction 518.3 11.9 298.6 13.9 Total services 1,470.8 33.9 869.0 40.5 Total 4,337.6 100.0 2145.5 100.0 Source: Armenia Social Accounting Matrix (SAM) 2002. See Part 2 of this book. Note: The above table displays Armenia's production and trade statistics for 2002. The larg- est single output sectors were construction, trade, and processed foods. Other large activities include the public sector, cut gems, transportation services, and electricity output. Food is probably more important to the Armenian economy than it appears in this table. If we com- bined the three largest food industries (processed foods, dairy, and meats and grains), then consumable food would be the largest single industry in Armenia. The total combined output for these four industries accounts for 25.1 percent of total production. If we reorganize these statistics according to value added, a different story emerges. Value added is generally high in the nontradable sectors such as construction, trade, banking, and public services. Food and food products (crops, vegetables, and fruits) are the leading tradable sectors in the economy. Growth in these sectors is more likely to contribute to higher personal income for Armenians in the short term. Other manufacturing sectors such as machinery and gem cutting have Supply to domestic market Exports Import Share in Share in Share in US$ output US$ output US$ Armington supply 155.2 99.8 0.3 0.2 50.7 24.6 126. 9 95.8 5.6 4.2 24.0 15.9 15.8 100.0 0.0 0.0 14.7 48.1 98.1 95.2 5.0 4.8 19.9 16.9 342.1 99.9 0.5 0.1 6.8 2.0 158.9 99.9 0.1 0.1 23.1 12.7 75.7 100.0 -- -- 156.7 67.4 27.2 41.3 38.7 58.7 1.1 4.0 379.6 88.1 51.1 11.9 46.7 11.0 77.3 95.6 3.6 4.4 30.0 28.0 23.5 44.7 29.1 55.3 40.2 63.1 44.7 84.4 8.3 15.6 81.8 64.6 11.3 4.2 258.0 95.8 206.4 94.8 65.3 59.4 44.7 40.6 55.3 45.9 33.0 37.2 55.7 62.8 185.7 84.9 56.1 91.8 5.0 8.2 52.1 48.2 152.0 100.0 -- -- 0.0 0.0 149.8 91.8 13.4 8.2 5.6 3.6 512.1 98.8 6.2 1.2 2.8 0.5 172.2 66.3 87.7 33.7 24.0 12.2 447.2 100.0 -- -- 0.0 0.0 44.9 87.0 6.7 13.0 11.3 20.1 354.8 97.9 7.7 2.1 7.8 2.2 23.1 24.6 70.6 75.4 60.5 72.4 92.6 100.0 -- -- 0.0 0.0 897.0 98.7 11.5 1.3 139.3 13.4 690.9 66.7 455.5 43.9 698.3 50.3 254.9 63.2 38.7 9.6 157.8 38.2 512.1 98.8 6.2 1.2 2.8 0.5 1,284.7 87.3 186.1 12.7 109.2 7.8 3,639.6 83.9 698.0 16.1 1,107.4 23.3 the lowest share of value added in production, which implies that these industries comprise mostly intermediate inputs and that the short-term spillover effects from these sectors are insignificant. These sectors have, however, the highest export potential and have considerable implications for the medium-term and long-term growth of the Armenian economy. Some industries, such as cellular telecommunications, mining, or utilities, may exhibit a high degree of value added in production, but since those industries are capital-intensive, their contribu- tion to personal income will depend upon who collects the capital rents. For example, a for- eign-owned mine may capture significant rents, but most of this income could be forwarded offshore. The table above also displays exports, imports, and total domestic supply for each sector. Import and export volume is dominated by a single sector: uncut and cut gems. This sector alone represents about one-fourth of total trade volume for Armenia. Export potential is highest in the manufacturing sector, as most of the manufacturing subsectors show high export/output shares. 38 Part I: Policies to Sustain Economic Growth Figure 2.1 Progress in First-Generation Reforms as Revealed in Values of the EBRD-Based Aggregate Index in 1992­2004 in Selected Transition Economies 100 90 80 70 60 50 percent 40 30 20 10 0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 year Lithuania CEEC-8 median Armenia SEE-4 median Source: Authors' calculations based on data from EBRD Annual Transition Reports. Note: The aggregate is the average of scores ranging between 1 (no liberalization) and 4.5 (liberalization at the levels of highly developed market economies) for prices, for- eign trade and exchange rate regime, and small privatization. It has been normalized with 1=0 and 4.5=100. SEE-4 includes Albania, Bosnia and Herzegovina, Macedonia, and Serbia and Montenegro. and large enterprises, and the reversion of land to private ownership. The experience of advanced transition countries (CEEC and the Baltics) suggests that the following factors are important for sustained rates of growth: undertaking deep structural and institutional reforms cen- tered on creating contestable and competitive conditions for economic activity, entrenching property rights, facilitating banking and capital markets to provide low-cost intermediation, fostering an environment for innovation and product diversification, and locking in integration with international services and factor markets. Constraints to Sustained Growth 39 In identifying the constraints to the maintenance of high growth rates, the framework building on the idea of growth diagnostics3 can be useful. This permits a focus on the country-specific constraints to growth as the first step toward identifying a growth strategy for the country. The question to be answered is not so much, "What creates success?" but rather, "What are the constraints to better performance?" The work on growth diagnostics suggests that constraints could be shown as a tree diagram (see Figure 2.2). The driving factors of growth are seen as the private return to accumulation (that portion of the social return which is subject to private appropriability) less the cost of financing accumulation. In Armenia, observation indicates that social Figure 2.2 Challenge of Sustaining and Diversifying Sources of Growth toward Those with Stronger Spillover Effects Macroeconomic stability (good progress) High social Low private High cost returns appropriability of finances Productivity gains Competition framework Low domestic savings (satisfactory progress) weak (underlying vulnerability) (binding constraint) Investment in Weak bank corporate infrastructure Poor contract governance (satisfactory progress) enforcement (structural weakness) (binding constraint) Stable price, fiscal, trade Insecure regime for regimes Corruption and arbitrary secured loans (good progress) state behavior (structural weakness) (structural weakness) Scientific and entrepreneurial Burdensome cost of endowment backbone services (satisfactory progress) (telecom, transport) (binding constraint) Diaspora and emigrant support Outmoded investment in (good progress) human capital (structural weakness) Source: World Bank staff. 40 Part I: Policies to Sustain Economic Growth returns are high, based on the rising productivity of factor use, but pri- vate appropriability is compromised by poor property rights protection and enforcement, inflexible labor regimes, and arbitrary customs and tax administration. The cost of finance remains high, partly because of low domestic savings rates but more fundamentally because of barriers to intermediation related to secured transactions and extremely weak corporate governance. The evidence suggests that the high growth rates of the past decade have resulted from productivity gains in the use of labor, with little additional labor use. Evidence for weaknesses in private appropri- ability and inadequacies in the competition framework comes from the large degree of informality in the economy (discussed in detail in Chapter 3), the use of informal mechanisms for the protection of prop- erty rights, and the perpetuation of monopoly positions in produc- tion and distribution. Evidence for the intermediation constraint arises from the internalization of finance through firms and business groups, the high shadow price of capital, and the dependence of investment on external financing. EVALUATING THE STATE OF THE BINDING CONSTRAINTS While progress in first-generation reforms (such as free price forma- tion, the convertibility of domestic currency for current account trans- actions, liberal trade and investment regimes, and private ownership) can be readily assessed, this is not so with second-generation reforms covering, as a rule, actions, policies, and institutions where progress can be gauged, at times, only indirectly. Moreover, the binding con- straints are usually highly intertwined. Progress in one area can be neutralized by the lack of progress in another area. For example, large- scale privatization may be completed, but its benefits may not materi- alize if privatized companies are subsidized by still state-owned banks (the situation in the Czech Republic in the mid-1990s), or if competi- tion policy is too weak. Excessive minimum capital requirements or instabilities in tax or customs policy or administration may offset the benefits for business formation and growth offered by the simplicity of procedures for business entry. Figure 2.3 compares structural reform progress in Armenia to three other country groupings. The figure uses a structural reform index (SRI), which was derived from EBRD transition indicators for com- petition policy, government and enterprise restructuring, large-scale privatization, banking reform and interest rate liberalization, and policy and regulations for security markets and nonbank financial institutions.4 Constraints to Sustained Growth 41 Figure 2.3 Progress in Structural Reforms: Values of SRI, 1994­2004 70 60 50 40 percent 30 20 10 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 year Lithuania CEEC-8 median Armenia SEE-4 Source: Authors' calculations based on data from EBRD Annual Transition Reports. Note: The aggregate is the average of scores ranging between 1 (no liberalization) and 4.5 (liberalization at the levels of highly developed market economies) for govern- ment and enterprise restructuring, competition policy, banking reform and interest rate liberalization, security markets and nonbank financial institutions, and large-scale privatization. It has been normalized with 1=0 and 4.5=100. SEE-4 includes Albania, Bosnia and Herzegovina, Macedonia, and Serbia and Montenegro. The results presented in Figure 2.3 show that moving to the insti- tutional environment of highly developed countries takes time. The median value of an SRI for new EU CEEC-85 member states stood in 2004 at 66 percent of the level of mature market economies. While this figure is above the level of institutional maturity achieved by Armenia (40 percent), the CEEC-8 countries have a long way to go to catch up with the highly developed countries. So does Armenia. The greatest "distance" (Table 2.1) separating it from institutions in Lithuania is in competition policy (at 29 percent of the "normal" desired level) and in security markets and nonbank financial institutions (29 percent). Competition policy and financial 42 Part I: Policies to Sustain Economic Growth intermediation are key binding constraints. Other institutional areas where the SRI index is below average for second-generation reforms are government and enterprise restructuring (38 percent) and banking reform and interest rate liberalization (38 percent). The most advanced area of structural reforms is large-scale privatization (67 percent). Another interesting observation that can be derived from data in Figure 2.3 is that Armenia's reforms have moved in leaps that extended over three to four years and that they appear to have slowed in the areas of competition policy, financial sector reforms, and governance over the past three years. The critical weaknesses in competition deserve particular attention. Although reforms and large investments in infrastructure have gen- erated a strong short-term growth response, the creation of a strong monopoly in the area of communications--a strategic sector--damages competitiveness and future growth. Telecommunications infrastructure in Armenia is currently one of the least developed in the CIS, despite the fact that it was among the earliest of the CIS to privatize. The 15- year monopoly granted to a private investor has led to the extremely high cost and poor quality of telecommunications services. Research shows the large and often economy-wide impact of the cost and diver- sity of telecommunications services on investment, international inte- gration, and growth. Competition in the backbone sectors remains restricted. Competi- tion forces tariffs down and boosts the quality of services, thus facili- tating progress in all other sectors. It also encourages utilization of a skilled, technically qualified labor force, which has traditionally been a comparative advantage for Armenia. Competition in back- bone services has strong spillover effects on household welfare and employment through multiple channels. Competition and transpar- ency in backbone services can also stimulate private investment (both domestic and foreign) and can entail technology transfer and techno- logical capacity building, as well as human resources development. (Box 2.2 discusses gains to be made from the liberalization of the service sectors.) The quality of service links is critical in the location decisions of multinational corporations and in a country's capacity to attract out- sourcing activities. Services linking production operations include procedures for the simplification and harmonization of international trade; the state of infrastructure and its management; and the provi- sion of such backbone services as telecommunications, banking, insur- ance, transportation, and business services. Together with customs, related border clearance regulatory procedures, technical standards regulations, and port efficiency, service links shape the ease and speed with which goods and services move across national borders and, Table 2.1 Progress with Second-Generation Reform, 1994­2004 Country 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Large-scale privatization Armenia 0 29 57 57 57 57 57 57 67 67 67 Lithuania 57 57 57 57 57 57 57 67 76 76 76 Government and enterprise restructuring Armenia 0 29 29 29 29 29 29 29 38 38 38 Lithuania 29 29 57 48 48 48 48 48 57 57 57 Banking reform and interest rate liberalization Armenia 0 29 29 38 38 38 38 38 38 38 38 Constraints Lithuania 29 57 57 57 57 57 57 57 57 57 57 Security markets and nonbank financial institutions Armenia 0 0 0 0 29 29 29 29 29 29 29 Lithuania 29 29 29 38 38 48 57 57 57 57 57 to Competition policy Sustained Armenia 0 0 0 0 0 0 0 29 29 29 29 Lithuania 29 29 29 38 38 38 48 57 57 57 57 Note: The EBRD transition scores ranging between 1 (absence of institutions supporting competitive markets) and 4.5 (presence of institutions at Gr the levels of highly developed market economies) are normalized with 1 = 0 and 4.5 = 100. The difference between 100 and the actual value repre- owth sents the distance to the level prevailing in highly developed market economies. Source: EBRD (various years). 43 44 Part I: Policies to Sustain Economic Growth Box 2.2 Gains from the Liberalization of the Services Sectors The importance of backbone services such as finance, telecommunications, and transport is threefold. First, these services constitute major inputs into the production of goods and services--including agriculture--and their cost often accounts for a major share of the total cost of production. Con- sidering that services contribute on average around 10­20 percent to the production cost of a product and account for all trading costs (includ- ing transport, trade finance, insurance, communications, and distribution services), the savings from liberalization of the services sector can be sub- stantial (Hodge, 2002). Second, services related to education, training, and health influence the productivity and quality of human capital. Thanks to lower input costs and higher quality inputs, these services help industries and agri- culture improve their capacity to compete with imports. Last but not least, cheap backbone services combined with efficient and unobtrusive customs is necessary for domestic firms to participate in the emerging division of labor based on international outsourcing, just-in-time production, and supply-chain management. The poor qual- ity of backbone services and trade facilitation deters foreign firms from incorporating domestic firms into their supply chains and also acts as a barrier to other types of trade. This so-called supply chain network has been the fastest growing component of world trade over the last two decades, generating employment and assuring access to most dynamic markets. The liberalization of the services sector and its opening to external competition will improve the quality of these services and reduce their prices. Empirical research suggests that liberalization may generate gains that substantially exceed those that come from merchandise trade liber- alization undertaken in isolation. Cross-country growth regressions for ECA economies, juxtaposing EBRD transition indicators and per capita GDP growth, point to a strong, linear, positive relationship between service sector policy reform and per capita GDP growth (World Bank, 2005b). The econometric study of the impact of liberalization in basic telecommunications on sectoral performance in 86 countries has found that complete liberalization has paid off in terms of higher teledensity (8 percent higher than in countries following the route of partial liberaliza- tion) and labor productivity. See Fink, Mattoo, and Rathindran (2002). therefore, are crucial to trade in goods. The weakest link in the "serv- ice chain" in a country may tip the balance against including a firm in the global supply chain. As a consequence, improvements in the domestic business climate may limit positive economic effects if they are not accompanied by similar improvements in trade facilitation. In Constraints to Sustained Growth 45 short, weaknesses in any of these services may lead to the exclusion of a country by firms searching to outsource activities abroad. Finally, it may be noted that the indicator of progress in structural reforms appears also to provide a good measure of the quality of governance and the incidence of corruption. The value of the correla- tion coefficient for the values of the SRI for 27 transition economies in Europe and Central Asia and the Corruption Perception Index (CPI) compiled by Transparency International was 73 percent in 2004. The positive correlation is higher for governance indicators (which are an average of political stability, government effectiveness, and regu- latory quality)6 as compiled by the World Bank. For the 1998­2002 averages, its value was 93 percent. In other words, countries with higher values for the structural progress indicator are better governed and are less corrupt. SUMMARY Armenia's strong growth performance is due to the completion of first-generation structural reforms and the persistence of the condi- tions for macro-stability. In these respects, Armenia's growth perform- ance is much closer to the record of the CEEC-8 countries than to the CIS countries or the Balkan nations. Future opportunities for growth will depend on the pursuit of second-generation reforms: competition policy, banking and capital markets, deeper international integration, and the laying of the base for innovation and knowledge. The chief binding constraints to growth are the weak conditions for fair com- petition and the shallow nature of financial intermediation. A strong monopoly in telecommunications raises costs, and restrictive air trans- port policies also handicap the landlocked and blockaded economy; these are policy-induced failures. Such weaknesses prevent the emer- gence of strong supply chains, to the detriment of future growth. NOTES 1. Armenia's external debt management has improved over time. Bilateral debt to Russia has been settled through a number of debt-to-equity swap oper- ations, which improved Armenia's financial position but increased its depend- ence on Russia through the monopoly gas suppliers Gasprom and Itera. 2. These three policy areas constitute the core of the first-generation reforms. Scores for each area range between 1 for no reforms and 4.5 for reaching the average conditions in highly developed economies. These scores have been nor- malized for this analysis, with 1 = 0 and 4.5 = 100. The average of these three indicators amounts to the achieved progress in first-generation reforms. 46 Part I: Policies to Sustain Economic Growth 3. Hausman, Rodrik, and Velasco (2004). 4. The average of scores in each of these areas--normalized for a minimum score of 1 = 0 and a maximum score 4.5 = 100--is the indicator of progress in structural reforms. For countries that achieved institutional maturity in terms of establishing institutions supporting competitive markets, the indicator is equal to 100, whereas for those that have not started the process it equals zero. 5. The CEEC-8 are the Czech Republic, Estonia, Latvia, Lithuania, Hungary, Poland, Slovenia, and the Slovak Republic. 6. There are two arguments in favor of focusing solely on these indicators: First, these indicators are critical dimensions of a business climate. Political stability affects investment decisions and, in extreme situations, its absence may disrupt economic activities. The quality of regulation is of little relevance unless supported by the government's capacity to enforce regulations. Second, three other indicators pertinent to such dimensions of governance--the rule of law, the control of corruption, and voice and accountability--are not taken into account because they do not yield extra information. For all transition economies, they are strongly correlated with the selected three indicators, with the values of correlation coefficients equal to or above 0.9. For the expla- nation and data, see www.worldbank.org/wbi/governance/gov_data.htm. Chapter 3 The Framework for Competition W ithin the Armenian economy, competition is hobbled by weak- nesses in the law such as the competition law and the law on contracts. Moreover, the implementation of laws is undermined by the prevalence of informal institutions, such as in corporate govern- ance arrangements, by business regulation and governance rules and practices, such as arbitrary and unpredictable behavior of tax and cus- toms administrations, and by policy-induced high costs, such as tel- ecommunications and transport costs, that fall on the shoulders of all enterprises. The conditions for competition in Armenia diverge from the ideal of a level playing field, particularly in sectors related to trade and distribution, where monopoly profits are being earned and wel- fare losses are being imposed on the public. This chapter examines the interplay of these factors on the state of competition and the prospects for sustained growth.1 GOVERNANCE THAT HINDERS COMPETITIVENESS Given the intimate link between the governance standards and the level playing field essential for fair competition, this section provides an analysis based on tools developed by the World Bank to measure the quality of governance (Kaufmann, Kraay, and Mastruzzi 2003). As international experience shows, FDI inflows and competition move up with improvements in the quality of governance. Table 3.1 reports the values of three indicators of governance (political stability, government effectiveness, and regulatory quality).2 A single aggregate index of gov- ernance is a simple average of the values of these three governance indicators. Their values range between ­2.5 (the worst case) and +2.5 (the best case). 47 48 Part I: Policies to Table 3.1 Selected Indicators of the Quality of Governance, 1998­2004 Sustain Political stability Government effectiveness Regulatory quality Aggregate (average) Economic Countries 1998 2000 2002 2004 1998 2000 2002 2004 1998 2000 2002 2004 1998 2000 2002 2004 Armenia ­0.5 ­0.7 ­0.5 ­0.5 ­0.5 ­0.9 ­0.4 ­0.3 ­0.5 ­0.4 0.1 0.1 ­0.5 ­0.7 ­0.3 ­0.3 Lithuania 0.4 0.4 0.9 0.9 0.2 0.4 0.6 0.7 0.2 0.5 1.0 1.2 0.3 0.4 0.8 0.9 Gr CEEC-8 median 0.9 0.8 1.0 0.9 0.5 0.6 0.7 0.7 0.8 0.6 0.9 1.1 0.7 0.7 0.9 0.9 owth CIS-11 median ­0.2 ­0.5 ­0.4 ­0.9 ­0.7 ­0.9 ­0.8 ­0.8 ­0.9 ­1.2 ­0.8 ­0.6 ­0.6 ­0.9 ­0.7 ­0.8 Source: Derived from data available at www.worldbank.org/wbi/governance/gov_data.htm. The Framework for Competition 49 The quality of governance in Armenia has improved significantly since 1998, although the pace slowed down somewhat in 2002­04. The value of the aggregate indicator of governance, after a significant dete- rioration in 2000 triggered by both political instability and a contraction in the state's capacity to govern, improved considerably in 2000­02 and registered a less pronounced improvement in 2002­04 (see Figure 3.1). The improvement over 2000­02 coincided with progress made in second-generation reforms as captured in EBRD structural indicators, where the largest push came during this period. Armenia scores high among transition economies in terms of the overall quality of governance. Only the CEEC-8, new members of the EU, and three other countries in the EU-accession pipeline had, in 2004, higher values in the aggregate quality of governance. Further- more, Armenia's lead over other CIS economies increased significantly in 2002­04. While the median value of the aggregate quality of govern- ance in CIS-11 also increased in 2000­02, trends diverged in 2002­04, albeit with Armenia recording a limited improvement (Figure 3.2). Figure 3.1 Aggregate Quality of Governance, 1998­2004 1.00 0.80 ­2.5 0.60 and 0.40 +2.5 0.20 0.00 between -0.20 -0.40 ranging -0.60 Index, -0.80 -1.00 1998 2000 2002 2004 year Lithuania Armenia CEEC-8 median CIS-11 median Source: Derived from data available at www.worldbank.org/wbi/governance/ gov_data.htm. Note: CIS-11 countries do not include Armenia. 50 Part I: Policies to Sustain Economic Growth Figure 3.2 The Average of Government Effectiveness and Regulatory Quality, 1998­2004 1.00 ­2.5 1.00 and 0.50 +2.5 0.00 between -0.50 ranging -1.00 Index, -1.50 1998 2000 2002 2004 year Lithuania Armenia CEEC-8 median CIS-11 median Source: Derived from data available at www.worldbank.org/wbi/governance/ gov_data.htm. Note: CIS-11 countries do not include Armenia. But much remains to be done. Although the quality of governance is higher in Armenia than in other CIS economies, it is significantly lower than in Lithuania or in a "median CEEC-8" economy. It is also significantly lower than in Bulgaria, Romania, and Croatia. In areas directly relevant to setting the level playing field essential for fair competition (government effectiveness and the quality of regula- tions), the average of these indicators is higher than the overall quality of governance. The greatest progress Armenia has achieved is in regulatory quality, the only area in which the value of the indicator is in a "positive" domain. However, decent levels in terms of regulatory quality, in the presence of a low capacity of the state to effectively implement policies and regulation, will not lead to significant welfare gains. Since better governance usually implies a lower level of corruption, the question is whether international investors--surveyed by Trans- parency International--have noticed the improvement. The point to The Framework for Competition 51 Table 3.2 Values of CPIs for Selected Countries and Groups, 1999­2000 and 2002­04 Index, 2004 Countries 1999 2000 2002 2003 2004 1999=100 Armenia 2.5 2.5 3.0 3.0 3.1 124 Lithuania 3.8 4.1 4.8 4.8 4.6 121 Median CEEC-8 4.4 4.2 4.4 4.4 4.6 105 Median SEE-5 2.5 n.a. 2.5 2.7 2.7 108 Median EU-15 7.7 7.6 7.3 7.7 8.4 109 Armenia in percent of CEEC-8 median 57 60 68 68 67 118 Source: Derived from data available at the website of Transparency International: http://www.transparency.org. n.a. Not applicable. be borne in mind is that the two conditions, although significantly correlated, are different, as governance also embraces the supply of public goods and services and their quality. The values of the CPIs provide information on international investors' judgment of corrup- tion in a country in a comparative context. International investors' perception of the Armenian investment climate has improved sub- stantially since 1998 (Table 3.2). While progress in reducing opportunities for rent seeking has not been impressive, Armenia has maintained its edge over other CIS economies. However, while the values of the CPI for Armenia in relation to CEEC-8 increased between 1999 and 2002, no perceptible improvement has occurred since then. Since the values of the CPI are low, indicating high levels of graft, the challenge is to introduce measures that will enhance transparency and simplify regulations. Considering fairly low levels of government capacity (as indicated by the value of government effectiveness indicator) in the implemen- tation and enforcement of laws and regulations, simplicity and clarity in the design of regulations is of importance. REGULATORY ENVIRONMENT WEAKNESSES THAT IMPEDE COMPETITION While the Transparency International CPI scores and the World Bank governance indicators capture either the perception of investors of the extent of graft in a country or an overall assessment of a country's 52 Part I: Policies to Sustain Economic Growth capacity to govern, the World Bank's Doing Business Survey captures tangible "legal" components of the investment climate together with the cost burden and time burden imposed on businesses across 145 countries (Djankov, La Porta, Lopez-de-Silanes, and Shleifer 2000).3 The survey is a useful diagnostic tool for conducting an initial assessment of the quality of the business environment, identifying most binding constraints, and assessing the overall ease/difficulty of doing business in Armenia in comparison with a selected group of countries (CIS and CEEC-10 economies). Armenia performs strikingly well in many meas- ures of regulation and the business environment, but nevertheless has a large informal economy, estimated to be in the range of one-third to nearly one-half of the size of the registered economy (Tunyan 2005). The next section of this chapter considers why this might be so, and attempts to drill down into the specific factors in the regulatory regime that constitute the binding constraints to fair competition. In particular, attention is paid to the interactions between the various characteristics of the business climate. Thus, for example, strong foreign trade facilita- tion measures will be of little relevance if the domestic environment discourages the entry of new businesses and encourages the exit of firms either for good or to the informal economy. Similarly, the limited capacity of firms to recover due payments from other firms discour- ages increased ties and limits the intensity of business activity, with negative consequences for domestic production and exports. THE EASE OF DOING BUSINESS IN TRANSITION COUNTRIES Table 3.3 presents the ranking of CIS and CEEC-10 economies in terms of the formal ease of doing business.4 Armenia ranks sixth among 20 transition economies. Doing business across all areas is on aver- age easier than in three new EU member states--Hungary, Poland, and Slovenia--and in all other CIS countries as well as Bulgaria and Romania. Only the Baltic states, the Czech Republic, and the Slovak Republic--all of them with strong reform credentials--score higher than Armenia. Armenia owes its high ranking to the procedural ease in hiring and firing employees, the simplicity and effectiveness of enforcing contracts, and effective procedures for closing a business. Its weakest areas are the registering of property, access to credit, and the large severance pay- ments that must be made when staff is fired.5 Overall, however, even in the most streamlined and reformed areas, the level is well below the best practice in transition economies. The gap is particularly large for labor market flexibility and registering property for Armenia. Table 3.3 Relative Ease of Doing Business in CIS and CEEC-10 Economies, 2005 Starting Flexibility of Registering Enforcing Closing Protecting Getting Ease of Country business employment property contracts business investors credit doing business Armenia 47 64 38 65 68 50 37 53 Azerbaijan 44 33 29 56 43 33 24 38 Belarus 18 50 48 54 44 17 44 39 Bulgaria 35 46 19 48 41 33 69 42 Czech Republic 33 56 28 72 18 100 89 57 Estonia 40 37 40 83 45 67 63 53 Georgia 39 51 18 54 53 83 56 50 Hungary 34 38 27 76 36 83 40 48 Kazakhstan 41 69 18 60 24 83 23 45 The Kyrgyz Republic 70 54 22 28 53 50 31 44 Latvia 48 30 15 77 100 83 70 61 Framework Lithuania 58 37 74 86 68 100 40 66 Moldova 33 50 26 51 41 50 36 41 Poland 32 49 19 51 60 67 57 48 Romania 79 17 17 51 27 33 53 39 for Russian Federation 43 69 21 62 77 50 13 48 Competition Slovak Republic 39 100 27 49 31 100 54 57 Slovenia 28 28 20 45 27 67 48 38 Ukraine 27 17 14 65 32 50 28 33 Uzbekistan 33 39 10 45 47 67 39 40 Worldwide 52 0 10 77 58 86 54 48 Source: Authors' calculations based on data in World Bank Doing Business database. 53 54 Part I: Policies to Sustain Economic Growth "REVEALED" EASE OF DOING BUSINESS The Doing Business Survey indicators measure the extent to which a country deviates from the best practice on each indicator within each area; they do not provide information on the comparative restrictive- ness of indicators within a country. An examination of the size of the informal economy throws light on such restrictiveness. Three causes of informality are important: the chances of getting credit, the fear of predatory administration, and a lack of trust in the state capacity to enforce contracts. While the survey captures both getting credit and contract enforcement, it does not capture informal administrative inter- ventions. Furthermore, the overall ranking procedure assigns the same weight to all areas of doing business. This book takes the average of the performance in terms of the informal economy and the ease of doing business as an indicator of the quality of the business environment. This indicator is referred to as the "revealed" ease of doing business. The large size of the informal sector, as is the case in Armenia, suggests that the hassle cost of doing business is higher than its high score would indicate. This observation applies also to other transition countries, as the value of a correlation coefficient between the values of the overall ease of doing business indicators and the respective size of the informal economy, though negative, is very low at 38 percent. This runs counter to the expectation that in countries with higher values of the aggregate "ease of doing business" indicator, the infor- mal economy should be smaller. In other words, the ranking in terms of formal overall ease of doing business diverges rather significantly from the ranking based on the size of the informal sphere. Further- more, while the ranking in terms of the size of the informal sector in transition countries is highly correlated with the incidence of cor- ruption as measured by Transparency International's CPI (correlation coefficient of 79 percent), the ranking in terms of formal ease of doing business is not so correlated; the value of the correlation coefficient is only plus 33 percent. None of the above comes as surprise. The ease of doing business indicators for CIS and CEEC countries cannot capture all of the factors determining the quality of the business environment. First, they have not been designed to do so, as they do not capture all variables rel- evant to the cost of doing business. For example, they do not take into account the fact that predatory tax administration operating according to ad hoc rules may dramatically increase the hassle cost of doing business, discouraging start-ups and encouraging exit from business activity. Second, the procedure used here to rank transition economies gives equal weights to all variables describing the formal regime of doing business. Countries with best practices on most counts may The Framework for Competition 55 rank relatively high in rankings despite weaknesses on other counts. For example, a small number of procedures, speed, and low admin- istrative fees to start a business would matter little in the presence of a huge minimum capital requirement. Since the size of the informal sector, which for transition economies happens to go in hand with the increase in the incidence of corruption, appears to be negatively linked to the friendliness of a business envi- ronment, a comprehensive assessment of the quality of the business climate should take into account both the formal ease of doing busi- ness and the size of the informal sector. One way of doing this is sim- ply to normalize the size of the informal sector in terms of its lowest level among transition economies (that is, the Slovak Republic with 18.9 percent, would be 100) and incorporate it into the final measure referred to as "revealed" overall ease of doing business.6 Table 3.4 presents the ranking in terms of revealed ease of doing business. The "fall" in Armenia's position, as well as in the position of other CIS economies, is a reminder that not only do other factors (not covered in the Doing Business Survey) determine the quality of the business environment but also that, within various areas, binding constraints exist that neutralize the positive impact of other changes. The ranking approximates well the progress achieved by the countries in building an institutional environment to support private business activity. It illustrates that in Armenia (as in Georgia and Russia), an apparently friendly regime is not sufficient to curb informality, an outcome that is detrimental to economic growth and productivity. The factors that countervail the strengths of the regime are discussed next. WEAKNESSES AND STRENGTHS OF FORMAL REGULATORY REGIME The formal regulatory burden of conducting business in Armenia is relatively low, especially when benchmarked against the CIS and CEEC-8 economies. Table 3.5 presents values of respective indicators for Armenia, the CIS economies, Russia, the CEEC-8, and New Zea- land. It also provides information on the best practice across the globe for each indicator. Indicators describe formal arrangements in terms of time and complexity in the following seven areas: starting a business, hiring and firing workers, registering property, enforcing contracts, closing a business, protecting investors, and getting credit. Armenia outperforms Russia in all indicators except the following: it is more difficult to hire in Armenia, it costs more to enforce a contract, and creditors recover less in terms of the value of a loan. Armenia rep- resents the best practice in the CEEC and CIS in two indicators: the 56 Part I: Policies to Sustain Economic Growth Table 3.4 The Size of the Informal Economy and the "Revealed" Ease of Doing Business Informal "Revealed" Difference economy, ease between percent of doing current of 2003 business Previous and Memorandum: GNI indicator ranking previous CPI, 2000­04 Armenia 46.3 46.8 6 5 3.0 Azerbaijan 60.1 34.5 18 2 1.9 Belarus 48.1 39.2 17 1 4.1 Bulgaria 36.9 46.4 13 ­1 4.0 Czech Republic 19.1 77.8 4 ­2 3.9 Estonia n/a 74.0 5 ­2 5.7 Georgia 67.3 39.3 7 10 2.1 Hungary 25.1 61.5 10 ­5 4.9 Kazakhstan 43.2 44.6 11 3 2.3 Kyrgyz Republic 39.8 45.8 12 1 2.2 Latvia 39.9 53.9 2 5 3.8 Lithuania 30.3 64.2 1 3 4.7 Moldova 45.1 41.4 14 2 2.3 Poland 27.6 58.2 8 ­2 3.8 Romania 34.4 47.2 16 ­6 2.7 Russian Federation 46.1 44.4 9 6 2.7 Slovak Republic 18.9 78.5 3 ­2 3.8 Slovenia 27.1 53.7 19 ­11 6.0 Ukraine 52.2 34.6 20 ­1 2.3 Uzbekistan 34.1 47.7 15 ­6 2.5 Source: Authors' calculations based on data in World Bank's Doing Business database and data from Web site of Transparency International. Note: CPI is the average for 2000­04. It assumes values between 1 (maximum incidence of corruption) and 10 (minimum). difficulty of firing index and the cost of closing a business (Table 3.5). However, neither indicator is necessarily an indication of the superior quality of the formal arrangements in the corresponding areas of hir- ing and firing workers and closing a business. Since the worst indicator does the most to discourage business activ- ity and thus contributes to the higher size of the informal sector under most circumstances, the following question is pertinent to an over- all assessment of the regulatory and policy hassle felt by businesses: which aspects of Armenia's regulatory environment are the most burdensome for private firms? While in each area Armenia's overall The Framework for Competition 57 score is high compared to its peers in the CIS and CEEC-8 countries, within some areas indicators point to weaknesses. These areas are labor markets (firing and hiring), getting credit, and contract enforce- ment. Contract enforcement takes more money than in CIS and CEEC- 8 countries (18 percent of the debt as compared with 14 percent), but other indicators favor Armenia. Despite its importance as a dimension of labor markets, Armenia's business environment has significant weaknesses even when bench- marked only against other CIS countries. The high value of the "dif- ficulty of hiring" index combined with an obligation to pay the wages of a fired worker over a period of 30 weeks reduces the flexibility of Armenia's labor market significantly. More favorable values for the remaining two indicators do little to lessen the constraints on busi- nesses in hiring when the demand for their product goes up and in firing when conditions deteriorate. Considering that research shows the crucial importance of labor market flexibility for both reducing unemployment and boosting investment, Armenia's poor performance in this area is a serious shortcoming. While the relationship between low unemployment rates and high flexibility is empirically firmly established, the impact of labor market flexibility on FDI appears to be significant as well. A recent study (Javorcik and Spatareanu 2004b) using the labor market indicators from the Doing Business database for 2002 shows that, all else being equal, flexibility in labor markets has a significant impact on FDI flows. For example, if the flexibility of the host country's labor market increases from the level of the Slovak Republic (inflexible prior to reforms in 2003) to the level of Hungary (flexible), the volume of investment goes up by between 14 and 18 percent. Moreover, FDI in the services sectors appears to be more sensitive to labor regulations than investments in manufacturing. Armenia also scores relatively low in terms of protection of inves- tors, with the disclosure index of 3 putting it on a par with Russia. The absence of regulations compelling disclosure has several nega- tive consequences: investments are lower as potential investors fear expropriation, the stock market is undercapitalized, economic growth is lower than under full disclosure, and returns from investment are lower. (World Bank 2005a, pp. 56­57). But more powerful financial constraints to doing business relate to other legal underpinnings of financial markets in Armenia. Although the cost of creating collateral is very low by both regional and inter- national standards, and the legal rights index is within a median range, there are countervailing deficiencies. The lack of market infor- mation on the quality of borrowers, combined with the absence of a 58 Table 3.5 Doing Business in Armenia and Selected Comparators in 2004 Best practice Part CIS CEEC-8 in CEEC New Best Area Indicator Armenia median Russia median and CIS Zealand practice I: Policies Starting a business Number of procedures 10 10 9 9 5 2 2 Time (days) 25 32 36 46 18 12 2 Cost (% of income per capita) 7.0 14.2 6.7 12.0 3.7 0.2 0.0 to Minimum capital Sustain (% of income per capita) 4.5 23.3 5.6 47.9 0.0 0.0 0.0 Hiring and Difficulty of hiring index firing workers (0­100) 33 33 0 33 0 11 0 Economic Rigidity of hours index (0­100) 40 60 60 60 20 0 0 Difficulty of firing index (0­100) 10 60 20 60 10 10 0 Gr Rigidity of employment owth index (0­100) 28 44 27 44 10 7 0 Firing costs: severance payments in weeks of wages)a 30 21 17 21 17 0 0 Registering property Number of procedures 4 7 5 5 3 2 1 Time (days) 18 71 354 72 3 2 1 Cost (% of property value) 0.9 2.2 9.5 2.1 0.5 0.2 0.0 Enforcing contracts Number of procedures 24 29 29 24 17 19 11 Time (days) 195 324 395 333 150 50 27 Cost (% of debt) 17.8 18.0 9.5 10.8 8.1 4.8 4.2 Table 3.5 (continued) Best practice CIS CEEC-8 in CEEC New Best Area Indicator Armenia median Russia median and CIS Zealand practice Closing a business Time of insolvency (years) 1.9 3.8 1.5 2.5 1.1 2.0 0.41 Cost (% of estate) 4.0 7.4 4.0 18 4.0 4.0 1.0 Recovery rate (cents on the US$) 39.6 5.3 48.4 39.8 85.0 71.4 92.0 Protecting investors Investors disclosure index (0­7) 3 3 3 5 6 5 7 The Getting credit Cost to create collateral Framework (% income per capita) 0.9 3.6 11.6 3.7 0.6 0.0 0.0 Legal rights index 0­10 4 5.0 3 6 9 9 10 Credit information index (0­6) .. 0.0 0 4 5 5 6 Public registry coverage (per 1,000 adults) .. 0.0 0 6 44 0 637 for Private bureau coverage Competition (per 1,000 adults) 0 0.0 0 17 380 978 1,000 Source: Data from World Bank (2005a). a. As noted, this cost will be greatly eased once the new labor code is put into effect. .. Negligible. 59 60 Part I: Policies to Sustain Economic Growth secured transactions framework, is a binding constraint on lending in Armenia. Movable property cannot be used effectively as collateral to secure loans, as repossession is time-consuming and costly. Neither can collateral-based lending function effectively in the absence of an effective framework for creating and enforcing claims. Banks price in the corresponding (large) risk into the lending spread. Armenia's disadvantage is lower in registering property: the number of procedures is only slightly higher (4 versus 3) but it takes signifi- cantly more time to complete the process (18 days versus 3 days). The same applies to closing a business. The time of insolvency is longer (1.9 years versus 1.2 years), although it is lower than in New Zealand (2 years), with the "best international practice" at 0.41 years. The recov- ery rate (that is, the percent of what creditors collect on their debt) of 39.6 cents on the U.S. dollar is below Lithuania's level, but similar to the CEEC-8 median of 39.8 cents. An indicator in starting a business area that is lagging behind comparator values is the cost of starting business, which is higher than in Lithuania but lower than the median value for the CEEC-8. Although Armenia appears to be a top performer among CIS coun- tries, it still has to neutralize the negative effects of being land-locked, as well as its remoteness to the centers of gravity of global markets. The frame of reference for Armenia's regulatory reforms should be at least the best practice in the CIS/CEEC-10 region, if not best inter- national practice. Compared to either reference point, the differences are huge, even though Armenia fares quite well overall compared to other former centrally planned economies. THE LINK BETWEEN THE FORMAL EASE OF DOING BUSINESS AND THE LARGE INFORMAL ECONOMY Surveys covering various areas of the business environment record a general dissatisfaction of businesses with the quality of regulations, administrative requirements, and bureaucratic behavior of state agen- cies, despite some improvements since 2002. The reasons for this assessment are manifold. The first barrier to conducting business arises from the difficulties encountered in interaction with tax administration, compounded by unstable and frequently changed rules and tax rates. This is perceived as a bigger problem than corruption by both local- and foreign-owned firms, and for the former it continues to be regarded as a significant bar- rier.7 Administrative Regulatory Cost Surveys (RCSs), which have been conducted annually in Armenia by the Foreign Investment Advisory Service (FIAS) since 2000, persistently identify tax administration as the The Framework for Competition 61 biggest obstacle. For example, in both 2003 and 2004 more than 80 per- cent of respondents identified tax administration as the largest obstacle. This view appears to be shared by foreign-owned firms to the tune of 90 percent of all respondents (FIAS 2003). Furthermore, foreign-owned businesses complained that the tax burden in 2003 increased consider- ably compared with 1999. Uncertainty associated with frequent changes in policy was flagged by around three-fourths of RCS respondents. The second barrier relates to the cost that businesses must incur in order to meet the requirements of the country's administrative regula- tions. According to the RCS, the costs remain huge and there has been no major improvement over the past several years. The costs are not related to formal and informal payments but above all to the burden in terms of time and resources that firms need to allocate to assure regulatory compliance. A corrupt bureaucracy applies regulations arbitrarily. The high level of corruption results in firms' directing activity underground in order to reduce their vulnerability to extortion by government offi- cials. Changes in legislation are only rarely announced or publicly disclosed before implementation. Bureaucratic procedures can be bur- densome and time-consuming when an investor negotiates a contract with the foreign government, as the contract may require the approval of several ministries. Widespread corruption continues to affect busi- ness and most often takes the form of bribery. According to a business climate survey carried out by the World Bank in 2004, 84 percent of companies surveyed were dissatisfied with legislative and administrative regulations in the country (World Bank 2005a). Armenia's business legislation in general is good, but imple- mentation of the laws has been poor, which has hampered business operations. The prevailing negative sentiment had worsened com- pared with previous years: the indicator stood at 63 percent in 2003 and 50 percent in 2002. The unequal treatment of companies by the authorities and unfair competition were the most widely cited impedi- ments to doing business. The third barrier is the practice by customs of extracting "irregu- lar" payments from businesses, a complaint that is corroborated by nontransparent procedures. Customs clearance procedures are applied with equal zeal not only to imports but also to exports, which further corroborates the opinions about customs erecting a nontariff trade bar- rier; generally, in other countries, customs efforts are largely directed only at imports. The time needed to complete customs clearance and the amount of effort undertaken by firms to deal with customs are shocking, especially for exports. Furthermore, the most obvious ben- efits of the computerization of customs services (that is, reducing the release time of consignments and slashing the documentation or 62 Part I: Policies to Sustain Economic Growth bureaucratic burden placed on a trader or a customs broker) have so far failed to materialize. Neither traders nor customs brokers have access to the system. The old practice of bureaucratic delays fueling corruption, combined with a lack of capacity in customs administra- tion, continues. Although Armenia is no exception to the generally poor record in most transition countries in establishing the rule of law and strength- ening the courts, the situation has improved on several important counts.8 For example, Armenia has made gains in judicial reforms. Armenia is "the only country where significantly more respondents viewed the courts as fair in 2002 than in 1999" (Anderson, Bernstein, and Gray 2005). However, considering that in 2000 only 2 percent of firms surveyed viewed judges as honest (CES 2000), the improvement should be measured from the low starting point. Fairness, however, is only one dimension of efficiently functioning courts. Other important dimensions from the point of view of conduct- ing business include the perception of the honesty or dishonesty of judges, the capacity to implement enacted commercial laws, and the ability to adjudicate disputes in an efficient and timely manner. It appears that firms regard Armenia's courts as relatively honest. In response to a question about the frequency of unofficial payments when dealing with the courts, firms surveyed under BEEPS identified the courts as not extracting bribes on a significant scale. The percent of firms negatively assessing courts in this dimension was among the lowest in the region--on a par with Slovenia, Estonia, and Lithuania. Moreover, Armenia, together with Lithuania, recorded the greatest improvement among 26 transition economies between 1999 and 2002. Although most respondents perceive courts in transition countries as neither honest nor fair, the percent of those that do not share this view is comparatively high at 30 percent, putting Armenia in the same group as Croatia and Latvia. Armenia, like most other transition economies, has experienced a growing gap between the enacted legislation setting the grounds for the market economy and the capacity of the courts to implement new commercial laws. This gap between the extent of commercial legis- lation existing in 1999 ("legal extensiveness") and the degree to which it was being implemented at that time ("legal effectiveness") was par- ticularly high for Armenia--well above the level in other transition countries (EBRD 1999). Although this "implementation gap" persisted in 2002, it has fallen sharply to levels similar to that in the Czech Republic. The ability of courts to adjudicate disputes in an efficient and timely manner reduces uncertainty in actual or potential business deals. Armenia, together with Hungary, the former Yugoslav Republic The Framework for Competition 63 of Macedonia, Latvia, Lithuania, and the Czech Republic, showed the largest improvement in the perceived speediness of courts between 1999 and 2002. While in 1999 only 10 percent of firms viewed courts as rapid in Armenia, this proportion rose in 2002 to 25 percent of respondents--the level comparable to Hungary. This proportion should rise further once the case management and court administra- tion systems currently developed are put in place. To summarize, the reasons for the large size of the informal sec- tor in Armenia include the high regulatory compliance costs due to official and unofficial payments, instabilities in state policies affecting the business sector, and predatory tax and customs administrations, rather than weaknesses in the judicial system. While efforts designed to reform the judiciary should continue, improving tax and customs administrations together with injecting stability into business regula- tions and policies stand out as major "other" ingredients accounting for the gap between the formal and revealed ease of doing business in Armenia. LAW AND INSTITUTIONS: FORMAL AND INFORMAL The state is not able to protect high levels of competition in domes- tic markets. Laws and institutions underpinning antitrust policies in Armenia provide a good illustration of the welfare and growth implications of the gap between the quality of regulations and their often opaque enforcement. Competition law is a recent innovation in Armenia. The law establishing the Commission for the Protection of Economic Competition (CPEC) was passed in 2000 and the CPEC itself began prosecuting cases in 2001. An analysis of the key aspects of the law shows the following: · The law prohibits collusion among enterprises but fails to distin- guish between vertical and horizontal collusion. Generally, econ- omists view horizontal collusion as being potentially much more damaging to consumers. Vertical agreements can also present problems, but the circumstances under which this occurs need investigation and analysis, case-by-case, within clear guidelines requiring the assessment of economic benefits and costs. Poten- tial problems could arise in Armenia because the technical capac- ity for this analysis with the CPEC is limited. · Article 4 of the law prohibits anticompetitive practices between loosely connected groups, although the criteria for determining such behavior are loosely set out and would require intensive investigation and several different levels of complexity. 64 Part I: Policies to Sustain Economic Growth · Article 6 of the law indicates that a firm may be considered hav- ing a dominant position in the market if it is not exposed to substantial competition, or if it commands a market share of more than one-third of total sales. Depending on the definition of markets, and the degree of concentration in the market under consideration, this threshold could be too low. Generally, public policy should be concerned with "conduct" and "performance" rather than with market structure as such. · The law forbids the abuse of dominant market power. In Arme- nia, the CPEC keeps a list of firms with such power and suppos- edly tracks them to ensure that their behavior does not involve abuse. Most of the cases brought by the CPEC have involved firms that are in this list. Cases have been prosecuted under Arti- cle 7 of the law, which applies to the abuse of dominant mar- ket position. FIAS has criticized this approach on the grounds that the CPEC appears to have an attitude that any firm on the list is automatically suspect, which essentially means any large company. Once a firm is on the list of dominant firms, scrutiny intensifies and transactions costs for the firm rise sharply. · The CPEC has brought a number of cases of "unjustified pric- ing" based on a methodology involving comparisons between the selling prices of the company and general price indices. This makes little analytical sense, in that there is no analysis of entry barriers or sunk costs or allowance for shifts in relative prices. Moreover, such a methodology implicitly assumes that profits or rates of return on investments should not differ between prod- ucts, companies, and sectors, regardless of differences in risk. · Article 8 of the law defines market concentration and control through mergers and acquisitions and revolves around a market share of 35 percent or more. Combined with Article 10 of the law, which prohibits concentration unless it fosters competition, Article 8 is used to justify action against such companies. · Article 9 requires that firms with substantial market power must be registered as being dominant. Consideration should be given by the government to bringing the competition law into line with the EU or U.S. model so that regula- tors and courts are better able to use well-established methodologies and precedents. Allowing private enforcement of the law can also contribute to efficiency. Implementation of the competition law presents a number of prob- lems. The weak institutional framework that exists in Armenia appears to result in some cases not being prosecuted, especially in the areas The Framework for Competition 65 outlined in later sections of this chapter, whereas there may exces- sive targeting of other firms. As in many other areas of competition in Armenia, the law itself is not the primary problem. The existing legislation, while inadequate in some areas, could be used to prosecute anticompetitive behavior. Rather, the problems lie in the application of the law, the protection that some groups appear to have, and the gener- ally weak private sector business environment that limits opportunities for new entrants. These issues are explored at greater length below. Arm's length contracting is a requirement of a well-functioning economic system. In Armenia, the law governing commercial transac- tions is inadequate, with the result that contracts tend to be informal and take place between people who know each other well. The effect is a barrier to entry and increased incentives for informal behavior, which is widely observed in Armenia. Problems with the law include the following: · Decisions on issues related to commercial transactions are based on an agglomeration of civil code laws, making the applicability of contract provisions uncertain. · Contractual agreements are based on "unless otherwise gov- erned by law" provisions rather than on "unless otherwise stated in the contract" provisions. Since many commercial transactions are subject to a wide range of laws rather than a more unified commercial code, this provision requires a full knowledge of all the laws and implementing regulations that may apply to an agreement, which compounds the problem of the unavailability of translations. · Not only does precedent not play a role in court decisions, but there is no record of cases being kept, although this is partly being addressed by a World Bank legal reform project. · In addition to the uncertainty arising from the commercial code, judges are widely viewed as less than competent or honest. As a result, businesses have little recourse if they are the victims of anticompetitive behavior. Laws, regulations, and formal institutions in corporate governance are particularly weak in Armenia. The EBRD ratings on corporate gov- ernance and enterprise restructuring place Armenia at the same level as Russia, but behind the Baltics and the Balkans. On several basic issues related to transparency, notably public access to the company registry and lists of founders and shareholders, Armenia is among the weakest of the CIS countries. The quality of corporate financial reporting also remains weak. 66 Part I: Policies to Sustain Economic Growth Moreover, while the securities law has some weaknesses, the great- est corporate governance deficiencies are found in the company law and other basic legislation affecting the corporate sector. The extent of such weaknesses also places an undue burden on government regula- tors. In the area of financial regulation, for example, the central bank is obliged to provide for extensive guidelines and proposed legisla- tion to make up for the shortfalls. Company legislation permits state- owned enterprises to get away with weak corporate structures, as seen, for example, in railways. Such limitations in the law and institutions for corporate govern- ance have serious implications for fair competition. Particularly in emerging markets, corporate governance has become recognized as an important part of the business climate. In virtually all of the CIS countries, weak corporate governance reinforces the dominance of a limited number of businessmen. At the same time, weak corporate governance undermines investor confidence in the capital markets, which as a result fail to provide adequate financing for capital-worthy projects and new enterprises. In the CIS countries, stock markets are littered with publicly traded companies that are little traded. Informal institutions and practices play an important role in deter- mining the fairness of the framework for competition. Many private sector participants in Armenia frequently make facilitation payments to ensure that dealings with the public sector proceed without undue delay. Unlike other countries, the system has so far not degenerated into massive corruption, although undoubtedly there is corruption that extends to the highest levels. Rather, it has evolved into one in which interlocking obligations arising from favors and interventions govern much of the interaction in the business community as well as between businesses and the legal system. A person who acts to inter- cede on behalf of another becomes a roof, and the benefactor incurs an obligation to return the favor in one form or another at some point in the future. In Armenia, there appear to be few rules which can- not be modified or adjusted through the intervention of a roof even within the judicial system. In sectors where a powerful roof exists, the incumbents appear to have the power to make life extremely difficult for new entrants. Foreign investors often do not have a network that would provide them with a roof and so frequently are more likely to have difficulty navigating their way through the various bureaucracies and the judi- ciary, as noted in all the surveys on foreign investment in Armenia. In addition, companies that are only exporters claim that they are singled out for unfavorable treatment, especially with respect to VAT refunds. The Framework for Competition 67 The roof phenomenon can act as a substantial barrier to entry in the economy for both foreign investors and new local businesses that attempt to compete with well-established companies and individu- als. Although there have been improvements recently, subtle methods are used to impose severe operating constraints on new competition, ranging from problems in clearing goods through customs to inspec- tions of various types by government officials. LARGE WELFARE LOSSES FROM WEAK COMPETITION IN PETROLEUM DISTRIBUTION As the economic effects of deficiencies in the competition framework for the utility (telecommunications) and transport (civil aviation) industries are discussed in subsequent parts of this book, this sec- tion considers the welfare losses imposed by lack of competition in the distribution of petroleum products. Having no oil or natural gas resources, Armenia is totally dependent upon imports of fuel prod- ucts, which constitute about 20 percent of total imports. This import group includes two major components in terms of volume: petroleum products (including gasoline and diesel) and natural gas imports. Statistics on petroleum product imports appear to have been under- reported. In spite of very rapid growth, gasoline and diesel imports into Armenia show that in 2000 there was a sharp drop in the import volumes and only very modest growth afterwards. This occurred in spite of a very sharp increase in automobile imports, which rough statistics indicate exceeded 14,000 vehicles per year, compared with 1,200 per year in 1999. Although oil prices rose in 2000, and demand may have been depressed by the fuel efficiency of recently imported automobiles and the conversion of some motor vehicles from gasoline to natural gas, it does not seem likely that in the face of such rapid growth fuel consumption could go down. The most likely explanation for the flat fuel imports is that there is smuggling of petroleum. The structure of market concentration underwent changes in the 1999­2000 period. In 1999 there was only one gasoline importer in Armenia, so that market power was located inside the borders of the country. In 2000 and 2001 the number of importers into the country increased; therefore, looking only at the number of operators within Armenia, the market could not be defined as uncompetitive. However, an analysis of the import chain reveals that there is only one supplier that provides the bulk (88 percent) of imported gasoline to importers. While the concentration of direct importers has declined, the virtually single sourcing of supply for gasoline importers means that the supply 68 Part I: Policies to Sustain Economic Growth chain remains highly concentrated. Even though it has been reported that one more company started to supply importers with gasoline in 2002, the share of the incumbent supplier firm remained dominant. It constituted about 70 percent of the total supply to importers. What is more interesting is that, even though the large number of retail companies operating in Armenia ensures retail competition, the retail mark-up price was substantially higher than the imported (wholesale) price (which we took to be the customs clearing price). The gross profit mark-up, most of which probably occurred at the whole- sale level, constituted more than half of the retail price in 2001 (Figure 3.3). Since competition at the retail level is strong, it is unlikely that excess profits are being earned by retailers--most of the profits prob- ably accrue to the wholesalers. These calculations (see Box 3.1) indicate that there appear to be large welfare losses arising from the concentra- tion of the petroleum market that could amount to the equivalent of over 1 percent of GDP. This loss figure could be even higher if we con- sider that there is a further profit margin earned by the single-source monopoly supplier of imports. This is not included in the calculations of the deadweight loss, which can occur as a result of the monopoly power of the one supplier of importers. This loss figure could also be higher if we also consider that the same situation dominates in similar markets (such as the diesel market). Figure 3.3 Gasoline Price Structure for the Period 1999­2001 (drams per ton) 100 146,833 225,207 203,376 90 80 70 60 50 percent 40 118,300 159,200 45,100 30 20 10 0 1999 2000 2001 year margin over wholesale price customs clearing price Source: NSS. The Framework for Competition 69 Box 3.1 The Petroleum Distribution Market To stress the negative impact of noncompetitive behavior in the gaso- line market we made a simple calculation of welfare losses (Table 1). (For the explanation of the terms used in the analysis below one is referred to Tirole 1988). We assumed a linear market-specific demand curve and estimated its slope and intercept considering price-increase v.s. demand-decrease scenarios during 1999­2000 (such a decrease in demand is mainly explained by price increase in international oil prices in contrast to the observed non-price impact (such as increased imports of cars after 2000) on demand during later years). Welfare Calculations in the Petroleum Market 1999 2000 2001 Retail Price (drams per ton) 265,000 370,000 362,500 Customs Clearing Price (drams/ton) 118,300 145,100 159,200 Quantities Consumed (tons) 258,100 181,400 187,500 Total Profit (drams, billion) 37,9 40,8 38,1 Estimated Slope of the Demand Curve ­1.4 ­1.4 ­1.4 Estimated Intercept of the Demand Curve 626,000 626,000 626,000 Quantities under Customs Clearing Price (tons) 363,000 343,800 333,700 Deadweight Loss (drams, billion) 7,7 18,2 14,8 Deadweight Loss (US$, million) 14,5 33,3 26,1 GDP (in US$, million) 1,845 1,911 2,118 Deadweight Loss as a percentage of GDP 0.8 1.8 1.3 It turns out that welfare loss, as a result of non-competitive behavior in the gasoline market alone, was averaged on 1.3% of GDP. This loss figure can be even higher if we consider that there is a double mar- ginalization (we have not include in the calculations the deadweight loss, which can occur as a result of monopoly power of one supplier of importers) and that the same situation dominates in many other similar markets (such as in diesel market). Moreover, the fact that the concentrated rather high gasoline market profits (about 3.7% of GDP) have not contributed to reducing unequal distribution of income in the country accentuates the importance of lib- eralizing the petroleum oils market. Source: NSS and authors' calculations. 70 Part I: Policies to Sustain Economic Growth ECONOMIC IMPLICATIONS AND POLICY CONCLUSIONS Lack of competition can be detrimental to efficiency over the long run. In addition, the apparent concentration of wealth among incumbents could reduce political support for reform and limit the benefits of rapid growth. There is ample evidence of poverty in Armenia, even in Yerevan; outside the capital it is far worse. Poverty reduction requires that the benefits of growth are spread to a substantial number of peo- ple, because even at current growth rates it will be some time before Armenia can afford adequate social safety nets. The more competition is restricted and activities are reserved for incumbents, the less likely it is that the benefits of growth will be widespread. There are other consequences of the dominance of incumbents in some sectors. The concentration of wealth provides substantial ability to grant "favors" to government officials to subtly (and sometimes, not so subtly) hinder competitors entering the market. The phenom- enon of roofs or protection rackets is one manifestation of the way in which incumbents maintain their market power. In other cases, it appears that there are direct attempts to prevent competition. Wide- spread subversion of the "rules of the game" by the rich and powerful sends strong messages to those who are on the outside. It contributes to a general attitude that taxes should not be paid, officials should be bribed, and goods smuggled. This does nothing to develop the foundation for a modern competitive economy. In the long run, promoting competition in Armenia is inseparable from promoting private sector development in the country. The ability to restrict competition frequently arises because the institutions that underlie private sector activity are underdeveloped, providing strong incentives for informal behavior. In Armenia, the public goods that pro- vide the foundation for private sector development are weak. The court system does not function effectively and arm's length contracting is risky, so that transactions tend to take place between those who know and trust each other. This gives powerful advantages to incumbents. Similarly, the financial system remains extremely underdeveloped even by the standards of low income countries.9 Hence, those that have substantial financial resources are in an strong position to maintain and strengthen their market dominance. The development of institu- tions supporting the private sector is the only long-run solution to promoting a competitive market environment. Legal reform and the reform of the secured transactions framework are two measures that will greatly enhance private sector activity. In addition, government institutions need to support the private sector rather than being the instruments for maintaining anticompetitive behavior that they are The Framework for Competition 71 now. In particular, reform of the customs service is the key to promot- ing competition. Should the CPEC be one of the pillars of a policy to promote compe- tition in Armenia? There are concerns that it could become yet another layer of bureaucracy that leads to corrupt "inspections," harassment of entrepreneurs, and protection of the incumbency. This concern can be addressed by emphatically assigning to the CPEC as its core role, within a strengthened statutory framework, the implementation of competition laws and the duty of advising the government on regula- tory reform. Arguments in favor of strengthening the CPEC recognize that it could be one way to address the restrictions on competition that abound in several sectors of the economy. The view of this book is that in its present form the CPEC is not a significant bulwark against anticompetitive behavior.10 It does not have the necessary skills, staff, or facilities to operate effectively. Although it has been successful in a limited number of cases, it is not a force in areas where competition is obviously restricted. Its report on the petroleum market is deficient in terms of depth as well as analysis.11 This does not mean that attempts to assist the CPEC should be abandoned. Over the longer term it could be one of the instruments for promoting competition in Armenia. Modest technical assistance is warranted to help the CPEC to better implement the law and improve its ITC resources. It is unrealistic, however, to perceive it as useful in the shorter term. The CPEC needs several years to evolve into any sort of instrument against restrictions on competition. Nevertheless, its continued existence is assured and the interests of the private sec- tor lie in its improving its technical competence. SUMMARY AND KEY RECOMMENDATIONS Governance, though improving, remains a hindrance to competition, particularly in the areas of government effectiveness, quality of regu- lations, and excessive discretionary behavior in the tax and customs administrations at the expense of rules-based discipline. Armenia does well in a CIS comparison, but lags far behind good interna- tional practice in setting a fair and attractive competition framework for investors. The large informal economy is illustrative of the high costs of doing business (especially taking into account the "hassle" factors) and of high labor (payroll) taxes, as well as weak incentives to going formal, such as the possibility of being able to obtain credit from banks. Formal laws and institutions addressing competition and 72 Part I: Policies to Sustain Economic Growth corporate governance are weak, enforcement is erratic, and informal barriers to competition abound. · Strengthen the competition law by clarifying collusive behavior and dominant market practices, and by modernizing the basis for judging pricing practices. · Enhance the administrative capacity of the competition commis- sion to review cases and enforce its rulings. · Modify the commercial transactions law to strengthen safe- guards. NOTES 1. Background material supporting the analysis of this chapter can be found in Chapters 10 and 11 of Part II of this book. 2. As has been mentioned in Chapter 2 (note 13), there are two arguments that focus solely on these indicators. First, these indicators are critical dimen- sions of the business climate. Political stability affects investment decisions and, in extreme situations of its absence, may disrupt economic activities. To be relevant, the quality of regulation must be supported by government capac- ity to enforce regulations. Second, three other indicators pertinent to dimen- sions of governance--the rule of law, the control of corruption, and voice and accountability--are not taken into account and do not yield extra information. All transition economies are strongly correlated with the selected three indica- tors, with the values of correlation coefficients equal to or above 0.9. 3. It provides data on the tangible components shaping the cost of conduct- ing business in Armenia and allows casting them in a comparative perspective. The survey covers the following eight areas: starting business, flexibility of employment, registering property, contract enforcement, doing business, pro- tection of investors, getting credit, and ease of doing business. Each area has several different indicators ranging from 1 (protection of investors) to 5 (getting credit and labor market flexibility) (see http://www.doingbusiness.org/) 4. The procedure used to assess the ease of doing business in Armenia is as follows. In order to assess Armenia's relative position among transition economies in terms of overall ease of doing business, we aggregate seven areas identified in the Doing Business Survey. We identify best practice indi- cators not worldwide but among transition countries. Indicators for countries are expressed in terms of percent of the best practice set at 100. We average indicators for each area covered by the survey, which yields an overall assess- ment of the ease of doing business for each country. 5. The new labor code in effect since 2005 greatly eases the severance pay burden for enterprises. The Framework for Competition 73 6. It is "revealed" because it takes into account the extent to which busi- nesses express a preference to opt out of the official economy or to stay in it. 7. According to the 2004 Regulatory Cost Survey, 71 percent of respondents identified corruption as a persistent problem, up from 66 percent over the previous year. Interestingly, corruption appears to be much less of a problem for foreign-owned firms than for domestic companies (FIAS 2003). 8. According to the World Bank study reviewing progress in judicial reform in transition economies, this progress has been much weaker in all transition countries than in other areas of institutional reform since 1990. Furthermore, firms' perceptions of the legal and judicial systems in transition countries are worse than comparable perceptions in most other regions of the world, according to a recent world wide survey of business executives (Anderson, Bernstein, and Gray 2005). 9. Discussed in the next chapter. 10. The commission has investigated a number of recent cases, among which was the successful elimination of a monopoly granted by the airport operator to a taxicab company. There have also been several hearings concern- ing the abuse of monopoly power by ArmenTel. The company had cut off the telephone lines of some Internet service providers without prior notification, suspecting them of providing voice-over-IP service in Armenia. 11. A four-page analysis concluded that there is competition in the distri- bution of petroleum because of the fact that there are several wholesalers/ importers within Armenia. The paper does point out that there is only one sup- plier/seller to the wholesalers but appeared not to view this as a problem. Chapter 4 Finance as a Barrier to Accumulation SHALLOW FINANCIAL INTERMEDIATION AND HIGH COST OF FUNDS Comparisons with CEEC and CIS countries show the shallowness of financial intermediation in Armenia (see Table 4.1). The financial system in Armenia remains small and bank domi- nated. A wave of bank failures since the end of 2000 has reduced the number of banks from 31 to 20. One foreign bank (a subsidiary of a large first-tier international bank) is a dominant player with about 18 percent of all assets and a quarter of all deposits. Total assets of the banking sector were a paltry 19 percent of GDP in 2004, substantially below the CIS average of about 30 percent in 2003. Broad money, which excludes foreign currency cash in circulation, remained flat at about 15 percent of GDP, compared with an average of about 20 percent for CIS countries. Foreign currency cash holdings are large and have been roughly estimated at one-third of GDP. The nonbank financial sector plays a negligible role in intermediation. The insurance sector is tiny, with an annual premium income of only about US$5 million. Since the savings to-GDP-ratio in Armenia is not dissimilar to those in countries with much deeper intermediation (such as the Baltics and Central Europe), the supply base of potential deposits for banks is clearly not a constraint. One explanation for the low intermedia- tion arises from the importance of remittances from abroad and the informality of the economy. The substantial inflow of remittances from abroad provides a source of funds to finance real estate and investment expenditures, which reduces reliance on the financial sector. Remittances reach a significant percentage of the population and provide individuals and small firms with funds for investment 75 76 Part I: Policies to Sustain Economic Growth Table 4.1 Financial Indicators in Selected Countries M2X/GDP Credit/GDP B/GDP (percent) (percent) (percent) Countries 1995­01 2004 1995­01 2004 1995­01 2004 Armenia 8.34 15.10 5.75 7.52 5.2 6.97 Azerbaijan 11.63 17.95 12.98 10.94 7.56 8.16 Czech Republic 67.83 66.63 69.75 41.44 21.57 10.36 Estonia 28.81 41.12 27.44 65.40 11.69 11.79 Georgia 6.67 13.27 4.9 18.83 4.96 8.40 Kyrgyz Republic 13.1 20.60 5.93 6.17 9.01 13.16 Latvia 25.16 38.27 12.22 51.29 11.94 13.01 Moldova 18.12 25.43 16.29 23.13 10.65 16.61 Poland 34.85 41.02 21.83 27.46 8.72 7.83 Russian Federation 17.15 31.63 11.02 25.94 7.68 14.21 Ukraine 12.58 36.36 7.94 25.67 6.99 12.27 Source: National central banks of countries presented in table. in small businesses and real estate purchases or improvements. The large informal economy, which is unofficially estimated to be at least one-third of GDP, also does not rely on the formal financial sector. The firms and individuals in this economy rely exclusively on cash for transactions, partly to evade taxes. This significantly diminishes the potential deposit base of the banking system and the market for products offered by banks and other financial institutions. But if deposit mobilization is weak, lending is weaker still. For example, the largest bank in Armenia, Hong Kong and Shanghai Banking Corporation (HSBC), lends a much smaller proportion of its liabilities (10 percent) to private business than it does in other coun- tries in which it operates (60 to 70 percent). Moreover, commercial banks hold large and growing excess reserves at the Central Bank, indicative of their comfortable liquidity positions even at low levels of deposit mobilization. Thus, deep structural problems lie at the heart of the abysmally low rates of bank lending to enterprises. Both the high cost of loans (averaging about 20 percent) and the lack of finance for business not only act as a deterrent to formality but also are a factor behind weak competition that reflects inefficiencies in the financial system. According to the 2003 FIAS RCS, more than half of the firms that participated in the survey believe that financing is the fourth leading constraint to the expansion of firms. Potential com- petitors cannot obtain finance, either in the form of working capital or for foreign trade. The result is that incumbents' market dominance is rarely threatened by new entrants and that wealth remains concen- trated among them. Finance as a Barrier to Accumulation 77 The rapid growth (albeit from the low base described above) in credit and monetary aggregates since 2004 is encouraging. The annual rate of growth of deposits has accelerated from 15 percent on aver- age in 2001­03 to 22 percent in 2004 (exceeding 50 percent in eight banks) and 18 percent in 2005. Similarly, the annual rate of growth of gross domestic nongovernment credit (excluding loans classified as losses) has accelerated from 24 percent in 2001­03 to 37 percent in 2004 (exceeding 65 percent in six banks) and 32 percent in 2005. IMPACT OF INSTITUTIONAL WEAKNESSES ON COSTS Banks have a limited capacity to assess risks and limit their lending to the insiders and activities with guaranteed returns (such as consumer lending) and require high and liquid collaterals. Some importers do use the banking system to raise letters of credit but the practice is not widespread. There is little export financing. The practice of using irrevocable letters of credit as a basis for providing funds to fulfill export orders appears to be nonexistent. As a result, the financial sys- tem cannot be said to support the growth of trade that, in turn, is the key to promoting more competition in Armenia. The essential skills in raising letters of credit are confined to a small number of banks. Before exporting under letters of credit can occur, training in the details of the procedures will be required for most banks in Armenia. Two indicators point to the limited degree of competition within the banking industry. The bank concentration indicator--the assets of the five largest banks as a share of total assets of the banking sys- tem--is given in Table 4.2. It points to a notable increase in bank concentration over the recent past. Moreover, out of these five banks, one large bank has about one- fifth of the total banking assets and almost a quarter of the banking sector deposits. Despite the recent increase in banking intermediation, the degree of concentration and the associated lack of competition among banks do not augur well for long-term financial stability. Second, the interest rate spreads remain high, also indicating con- centration and lack of competition (see Table 4.3); with very high Table 4.2 Bank Concentration Indicator, 2000­04 (in percent) 2000 2001 2002 2003 2004 Share of the assets of five largest banks in total assets of the banking system 46.8 48.4 54.1 54.5 56.3 Source: Central Bank of Armenia. 78 Part I: Policies to Sustain Economic Growth Table 4.3 Decomposition of Interest Rate Spreads, 1999­2004 Total bank percentages 1999 2000 2001 2002 2003 2004 1. II/L 28.32 20.70 17.24 12.96 12.30 12.29 2. IC/D 13.52 12.34 7.38 4.85 3.41 3.10 3. II/L­IC/D(spread) 14.80 10.17 11.91 9.87 10.47 10.78 4. Sum of factors below 14.65 8.87 8.46 7.21 7.95 7.76 5. NII/D(­) 5.01 3.62 4.72 6.17 6.06 5.76 6. NIC/D 10.42 8.38 8.11 8.18 7.69 7.38 7. Net prov/D 3.08 3.62 3.35 3.99 1.50 0.52 8. ROA A/D 3.89 ­1.16 0.34 0.18 3.84 4.88 9. II/L r 2.27 1.66 1.38 1.04 0.98 0.74 Residual () 3­4 ­0.15 ­1.30 ­3.44 ­2.66 ­2.52 ­3.02 Source: Central Bank of Armenia. Financial Sector Assessment Paper (FSAP), update team preliminary calculations. Note: Definitions used in the table: II--interest income of banks; L--assets ensuring interest income, IC--interest expenses of banks; D--interest-bearing liabilities of banks, II/L--derived interest income of banks, IC/D--derived interest expenses of banks, NII--noninterest income of banks; NIC--noninterest expenses of banks; Net prov.--net assignments to reserve funds on loans, securities, and receivables. ROA A/D--the ratio of total assets in cost-bearing liabilities weighted by ROA; ROA--return on assets; Residual ()--Residual emerges as we compare the statement data on income and expenditures of commercial banks (flow indicators) with the balance sheet indicators (supply indicators), and consider a simplified assumption that commercial banks invest the total nonreserved part of the attracted interest-taking liabilities into the interest- earning assets. nominal and real interest rates for loans being prevalent. The cost of interest-bearing liabilities has declined significantly and reached extremely low nominal levels that approximate to low single-digit rates in real terms in the recent past. The principal factors behind the high and rigid spread structure in Armenia are the poor state of corporate governance; weak institutions and practices that affect banking transactions, such as the availability and flow of information on borrowers; the ease with which secured transactions can take place; the costs experienced in contract enforce- ment, as well as the costs associated with the management of banks; and the high unit costs of bank activities arising from diseconomies of scale. These factors point clearly to the role that structural reforms can play in alleviating finance as a barrier to sustained economic growth. Banks' administrative and noninterest-bearing costs are very high; at the same time banks make significant earnings from noninterest- bearing transactions, which indicates heavy reliance on nonbank oper- ations. The incompetence of the bank management itself can become a risk to the system. First of all, it can support the growth of low-quality Finance as a Barrier to Accumulation 79 loans owing to the lack of capacity to implement proper credit analysis. Second, if the quality of financial services does not improve in line with the credit growth, this can "force" clients to use "nonbank alternatives" for financing. Table 4.3 substantiates the above description. The indica- tors provide evidence that the costs of the interest-bearing liabilities of banks have declined significantly over the last years. At the same time the earnings of interest-bearing assets declined at a lower speed. Analysis makes it clear that the main factors contributing to the higher interest rate spread remain the noninterest costs of banks. This indicates that the banking system suffers from inefficient bank man- agement and high administrative costs. These costs are even higher if we exclude the largest bank in Arme- nia, which perhaps indicates that the presence of this foreign-owned bank alone does not create competitive pressure on the domestic banking system to increase its efficiency and push the interest rate spreads down. Another factor contributing to the higher spread is the ratio of total assets in cost-bearing liabilities weighted by the return on assets. This indicator has high volatility due to the significant restructuring of the banking system over the last years. It shows a growing tendency over the last two years; although this indicates a profitable allocation of the low-cost resources, its rapid growth contains future risks for the banking system. It is only through greater competition, brought about by improve- ments in the lending environment, that spreads will decline sustain- ably. Since the cost of interest-bearing liabilities is very low, the fall in spreads will come through a lowering of lending rates. The key challenge for the authorities is to extend the recent deep- ening in intermediation and entrench confidence in the financial sys- tem while maintaining macroeconomic stability. Steps to be taken are summarized in Box 4.1. Rapid credit growth is a necessary element in the expansion of still shallow financial intermediation; moreover, such growth poses a limited immediate risk, as the banking system is highly liquid and well capitalized. Further steps are needed to improve the quality of financial intermediation and to increase the range and penetration of financial services. The rest of this chapter considers the policy reform tasks ahead. CORPORATE GOVERNANCE DEFICIENCIES AND THE COST OF FINANCIAL INTERMEDIATION Central to weak corporate governance is lack of transparent owner- ship and control of banks. The Detailed Assessment of Compliance with the OECD Principles of Corporate Governance indicated that, 80 Part I: Policies to Sustain Economic Growth Box 4.1 Priorities in Reform of Financial and Capital Markets ACTION Liquidity management Raise the stock of local currency monetary instruments. Lengthen the maturity of T-bills and bonds. Bank supervision and regulation Amend bankruptcy law for banks to extinguish bank owners' shareholder rights in case of insolvency. Revise central bank law to explicitly protect supervisors from civil liability. Strengthen bank corporate governance based on the proposal by the Central Bank. Enhance consolidated supervision, especially with respect to risk exposures of beneficial owners and affiliated companies. Insurance supervision and the insurance sector Complete regulations implementing the new insurance law. Corporate governance Strengthen coverage and enforcement of the company law, especially on fiduciary duties and accountabilities of boards. Improve accounting and auditing practices, financial reporting, and creditor rights and strengthen disclosure by beneficial owners. Pension reform Develop actuarial projections and analyze impediments to reform options before deciding on a pension reform strategy linking benefits to wages. Housing finance Focus reform on the development of a primary market for housing finance. Develop a liquidity facility to facilitate longer-maturity mortgage lending while ensuring that there are no implicit guarantees. Source: Based on Financial Sector Assessment Paper (FSAP). Finance as a Barrier to Accumulation 81 while company law governing open and closed joint stock companies is sound, implementation is weak. Only a few banks and corpora- tions choose to operate under this law, covered under the securities legislation and enforced by the exchanges and securities commis- sions. Many of the banks and corporations are closely held and have few owners, who do not see the benefits of adopting a more formal governance structure and who are not publicly traded. The enforce- ment of sound corporate governance practices through securities leg- islation and securities regulators is weak. Clearly, there is a need for a fundamental reform to enhance the coverage and enforcement of company law if corporate governance, disclosure, and accounting are to improve. Improvements in transparency and accounting are basic first-generation reforms in corporate governance. The opaque ownership structure of the corporate sector makes it more difficult for the banking sector to assess the risks of corporate- sector exposures. Official disclosures of direct and indirect ownership are substantially weaker than in other transition economies. No dis- closure of beneficial owners is required under Armenian law. This limits the ability of banks to ensure compliance with prudent lim- its on loan portfolio concentrations and related party transactions. Although legislation requires all companies to prepare their financial accounts in line with the Accounting Standards of the Republic of Armenia, in practice these standards are not observed or are imple- mented only partly due to lack of training. Steps should be taken to improve accounting and auditing practices and financial report- ing requirements and to strengthen public information; this should include beneficial owners of banks and publicly traded companies. Furthermore, the supervising boards of Armenian companies are not sufficiently effective or accountable. The company law should clearly establish the right of shareholders to decide on owner-manager rela- tions (in cases, for example, where a separation of company owners from company managers may be desirable and should be provided for), subject to the protection of the contractual rights of minority shareholders. Moreover, legislation should strengthen and clarify the fiduciary duties and accountability of boards of directors. The above-mentioned shortcomings in corporate governance can be addressed through legal, institutional, and supervisory reforms. Given the weak implementation of company law, the Central Bank has developed a proposal to accelerate improvements in corporate governance in the banking sector through legislative and supervisory means, and through the introduction of upward-consolidated supervi- sion to address exposures to beneficial owners. There are a number of specific reforms applying to the corporate sector that should be imple- mented to address the shortcomings in corporate governance. These reforms include improvements in accounting and auditing practices; 82 Part I: Policies to Sustain Economic Growth provision of financial reporting requirements; enhanced disclosure of the shareholders of corporations; improved access to information by the public, in particular by making the company registry publicly available; and the strengthening of the role of boards of directors. Such shortcomings in corporate governance detract from confidence in the banking system, lead to information gaps that raise the cost of loans through higher risk premia, and impose a "tax" that disadvan- tages local enterprises. The Central Bank has started to address gov- ernance, including the areas of accounting and audit requirements and transparency in reporting through regulations. However, a number of key weaknesses remain which, unless addressed, would add to risks and intermediation costs: Banks should reform corporate governance practices with a view to adopting the legal form of open joint stock companies to increase the transparency of ownership. The responsibilities and accountabilities of bank owners and boards of directors should be explicitly defined in law, includ- ing taking responsibility for the safe and sound operation of the institution; being informed of their banks' operating condition, in part through ensuring an adequate external and internal audit process; and providing accurate and truthful reporting to the public and the central bank. Central Bank remedial actions should extend to both boards of directors and significant participants in the case of unsafe and unsound bank operation and significant compliance breaches. In late 2005, the Central Bank adopted a plan for the strengthening of corporate governance in banks. Specifically, measures are envisaged to protect shareholder rights (especially those of minority sharehold- ers), to clearly define the powers of the boards and the executives of banks, delineate the fiduciary responsibilities of management, and publicize bank activities. Improvements in the legal and regulatory framework and proce- dures have strengthened the supervisory process, but some weak- nesses remain. It is encouraging that the Central Bank conducts a rig- orous supervisory process, with enforcement grounded in legislation, regulatory reporting, and prudential norms. But there are weaknesses. Bank corporate governance and risk management, including the eval- uation of risks in new banking products, requires further strengthen- ing. Furthermore, a framework should be established to strengthen the monitoring of the parent and affiliated companies of banks in the context of consolidated supervision. The Law on Banks and Banking should be amended to allow the Central Bank to reject bank licenses if the legal and managerial structure does not permit adequate identi- Finance as a Barrier to Accumulation 83 fication of ownership and control, thereby hindering effective supervi- sion. The dialogue between the Central Bank and the external auditors should be substantially improved, if necessary through legislation. In addition, the law should be amended to obligate external auditors to report to the Central Bank on issues of material importance arising from their audits. Communication among regulators should be estab- lished, and formal agreements for information sharing put in place. The Law on Banks and Banking should explicitly provide for protec- tion for supervisors from civil liability. CREDITOR RIGHTS AND SECURED TRANSACTIONS FRAMEWORK A recent National Bureau for Economic Research (NBER) paper used a sample of 129 countries to show that "stronger legal rights of creditors are associated with a higher level of development of private credit mar- kets" (Djankov, McLiesh, and Shliefer 2005, p. 16). The paper under- lines the fact that this is mainly true in developed countries where the legal framework and contracting rules are more developed. The protection of creditor rights is one of the main legal problems that hinder the growth of bank lending. The enforcement of collateral is difficult for banks and this, as we have already observed, translates into high lending rates. Although the legislation on allowing secured creditors to sell collateral without resort to a court is adequate, the enforcement and the infrastructures supporting the laws are very weak and underdeveloped and do not work in practice. Collateral enforcement is difficult because of the absence of a unified and digit- ally accessible registry for immovable items and title issues related to the ownership of property. The judicial capacity is weak owing to lack of experience, corruption, insufficient capacity, and the long lead time required to adapt to the new legislation. Measures are being imple- mented to improve the legal and institutional arrangements relating to secured credit, the registration of ownership, and security interests, and to reduce the inefficiencies in court systems that impede debt recovery, which should address most of the existing weaknesses in creditor rights. In Armenia, movable property cannot be used effectively as collat- eral to secure loans. There are numerous problems with the secured transactions framework. Floating pledges are not allowed, and there are no registries which allow pledges to be perfected. Repossession is time-consuming and costly. A particular problem with financing exports is that there is no provision which allows the use of future production as collateral, so that financing against export orders is not feasible. A similar problem arises with imports because there is no 84 Part I: Policies to Sustain Economic Growth provision for pledging goods that are not in the possession of the borrower; goods that are being imported, even though they have been paid for by letter of credit, cannot be seized in the event of default. The effect of the inadequacies of the collateral framework extends throughout the economy, with the result that banks correctly perceive lending as extremely risky.1 The stages necessary for establishing a framework for secured transactions are as follows: Creation. The process by which the creditor establishes a secu- rity interest in a specific property (the collateral) Priority. The process by which the lender establishes the priority of the security interest Publicity. The process that makes public the priority status of the security interest Enforcement. The process by which, upon the debtor's default, the creditor will seize and sell the collateral to satisfy his/her claim Each of these stages must function effectively for collateral-based lending to occur. Currently, in Armenia, none of them work well. In this context, efforts by the authorities to reform the framework for secured transactions are most encouraging. But a reform program is likely to require several phases, given the complexity and the multi- sectoral nature of the task involved. Comprehensive collateral reform requires that the whole process for the pledging of property be reviewed and changed. While rewriting the laws governing the ability of property to serve effectively as collateral is an integral part of this process, it is a long way from being all that is needed. Without a thorough revision of the whole system, collateral reform will remain elusive. There have been many attempts at such reform in other countries, most of them unsuccessful. Legal analysis, without corresponding analysis by economists of suggestions for reform, has failed in most countries in which it has been tried. Internationally known legal prac- tice may not represent best economic practice. In many places model laws have provided the foundation for attempts to reform the col- lateral framework. Thus, a good project requires close coordination among economists, international expert lawyers, local lawyers, and technical experts. Examples of successful reform do exist and serve as a model of what could be done in Armenia. A recent reform of the secured trans- actions framework in Romania has transformed Romania's lending environment by facilitating the use of collateral as security for lend- ing, not only from the banking system, but also from equipment sup- pliers, wholesalers, and agricultural suppliers. Before this reform there were many similarities between the Romanian financial sector and Finance as a Barrier to Accumulation 85 that in Armenia--severe financial underdevelopment, the inability of a large sector of the economy to access credit, and a distrust of banks. Furthermore, Romania is also a civil code country. Use of this model could have significant potential for the development of the Armenian financial system and could substantially reduce barriers to entry. The secured transactions reform in Romania tightly integrated diagnosis, drafting, and regulations, and this integration is one of the main rea- sons for its success. This methodology has rarely been followed in reform efforts elsewhere. Legislation before Parliament is intended to strengthen creditor rights. Specifically, measures are planned to make the seizure of col- lateral property and its disposition easier, to broaden the category of properties that may be used for collateral, and to abridge rights that may interfere with seizure of collateral. Overall, Armenia lacks a regularized system of credit that can be stimulated by mechanisms that provide efficient, transparent, and reli- able methods for recovering debt, including the seizure and sale of immovable and movable assets and the sale or collection of intangible assets, such as debt owed to the debtor by third parties. An efficient system for enforcing debt claims is crucial to an efficient functioning of the credit system in Armenia. Armenia's financial sector (possibly with help from the Central Bank and the Ministries of Finance and Justice) should promote an informal, out-of-court process for dealing with cases of enterprises with financial difficulties in which banks have a significant exposure. Armenia's credit-based economy requires predictable, transparent, and affordable enforcement of both unsecured and secured credit claims by efficient mechanisms. Although commercial transactions in Armenia have become complex as more sophisticated techniques are developed for pricing and managing risks, the basic rights govern- ing these relationships and the procedures for enforcing these rights have not changed much. Those rights that should enable parties to rely on contractual agreements, fostering confidence that fuels lend- ing, are not in place. Therefore, uncertainty about the enforceability of contractual rights increases the cost of credit to compensate for the increased risk of nonperformance. THE ROLE OF INFORMATION FLOWS IN REDUCING INTERMEDIATION COSTS The NBER study2 emphasizes that information sharing through public and private credit registries is a significant factor promoting intermediation. The authors also find that "in addition, public credit bureaus are strongly associated with private credit in the poorer, but 86 Part I: Policies to Sustain Economic Growth not the richer, countries, pointing to a possible role of government in facilitating information sharing" (Djankov, McLiesh, and Shliefer 2005, p. 26). The credit registry, which was introduced by the Central Bank in January 2003, aims to reduce credit risks by the creation of an infor- mation system on the creditworthiness of customers of banks and credit organizations operating in Armenia. However, while the credit registry has been beneficial to banks' supervisors, this information is limited to debtors of banks only. A private credit bureau began opera- tions in early 2004 and focused in its first year on obtaining infor- mation on debtors from the banks, other financial institutions, utility companies, and government offices. This information should improve creditors' ability to evaluate prospective borrowers' creditworthiness. However, the credit bureau has faced difficulties because most banks are reluctant to provide information for free and then to have to pay for services. Furthermore, based on tradition and culture, Armenian borrowers are reluctant to share their information with others, and ensuring privacy protection substantially constrains the effectiveness of the credit bureau. This credit bureau is currently revisiting its oper- ating model and expanding ownership to include banks with a view to remedying these obstacles. In countries where credit bureaus have been established success- fully, the central bank plays an important role in creating incentive mechanisms for commercial banks to share information on their cus- tomers. In Armenia, the Central Bank proved to be less than enthusias- tic, apparently because the credit bureau would compete with informa- tion that it sells. The lack of support by the Central Bank has, in turn, led to a waning of enthusiasm on the part of the commercial banks for the new credit bureau and only one has signed up for its services. In other countries in which credit bureaus have been established success- fully, it has usually been necessary for the central bank to insist that the commercial banks share information on their customers. Given the vital role played by the generation and transmission of information on creditworthiness in ensuring low intermediation costs, it is suggested that the Central Bank adopt an active role in facilitating private credit bureau development by the following means: Make credit bureau reports a requirement for lending coupled with mechanisms to ensure efficient entry and exit by bureaus to avoid the abuse of their exclusive rights. Require a borrower's explicit consent prior to anyone's gaining access to their files to ensure more privacy protection. Other actions that may prevent privacy violations include strict bank Finance as a Barrier to Accumulation 87 policies on gathering certain kinds of information, conditions on right to access, and validation of own files, and rules on the elimination of individual files after a certain period. Adhere to the following membership policies. Access to informa- tion by borrowers and credit institutions should be granted on the basis of membership. In other words, the principle of reciprocity should guide credit bureaus and should be stated in the contrac- tual agreement between the bureau and the credit institutions. Membership should not be based on fees of any form. However, credit bureau activity should be profit-oriented. In cases where a member provides inaccurate information or fails to provide data, sanctions (ranging from fines to loss of membership and hence denial to the bureau's files) should be imposed. Adopt a balanced ownership policy. Credit bureaus are exposed to potential conflicts of interest, especially if they are owned by a group of lenders: lenders want to exploit the information pro- vided by other lenders without disclosing their own. Therefore, Armenia should adopt a balanced ownership policy. For exam- ple, credit bureaus could be incorporated as private companies and owned by a consortium of lenders to create the incentive for information exchange. Alternatively, independent ownership, coupled with a proper membership policy, could provide incen- tives for lenders to exchange information. Regulate the quality of disclosed information. Apart from the efficiency debate and the coexistence of different institutional arrangements, there is concern regarding the optimal amount of information sharing and the content of the information. In contrast to relying on the general statement that information dis- semination reduces adverse selection problems (due to bad risks in the population of credit seekers) and makes the information on which the banks base their lending decision homogenous, the policy for the development of an efficient system for sharing credit information should be focused on the regulation of the quality of disclosed information. MONETARY OPERATIONS AND MONEY MARKETS While monetary policy has ensured low inflation, it has been con- strained by a low volume of securities for use in liquidity manage- ment; the stock of local-currency government debt is low at about 3 percent of GDP. As a result, the Central Bank must rely on foreign exchange operations as its primary market instrument to manage local 88 Part I: Policies to Sustain Economic Growth currency liquidity. This worsens the classic tradeoff between exchange rate volatility and a buildup of excess liquidity, which could pose an inflation risk. An increase in the supply of local currency securities would ena- ble monetary operations in local currency to concentrate on liquidity management. This would free foreign exchange operations to focus on managing large capital inflows. The recent agreement between the government and the Central Bank to increase the supply of securities will enhance the role of money markets. The monetary transmission mechanism is weak because banks are highly liquid, money market activity is low, and dollarization is high. Monetary transmission can be strengthened by the development of money markets, which should lead banks to move away from buy-and- hold investment strategies, to hold lower excess reserves, and to rely to a greater extent on money markets. Improvements in monetary trans- mission should also result from a strengthening of the broader finan- cial system infrastructure, including corporate governance and creditor rights, which would contribute to the narrowing of lending margins and lead banks to become less liquid by reducing lending risks. Finally, an increase in reserve requirements from their current level of 6 percent to reduce banking system liquidity could be considered if it would be difficult for market instruments alone to absorb the excess liquidity. But this instrument (which is effectively a tax on banks) is a blunt one, and market- or price-based instruments are preferred. INSURANCE, PENSION, AND HOUSING FINANCE REFORM The new Law on Insurance adopted in 2004 is comprehensive, but may be ambitious in light of the current level of supervisory exper- tise. The authorities should finalize and implement key regulations to improve accounting, auditing, and financial reporting, and solvency and reserves. The oversight of reinsurance also needs strengthening. The supervisory authorities should align the regulations implement- ing the new insurance law to international best practice and imple- ment them as soon as possible. Key areas include the following: Licensing. Given the very small size of the market, the existence of 24 insurers seems inefficient. The increase in minimum capital levels will help address this. Accounting, auditing, and financial reporting. Accounting, audit- ing, and financial reporting practices in the market are weak, as is monitoring of insurers' market conduct for consumer protec- tion. Regulations in these areas should move the market toward international practices. Finance as a Barrier to Accumulation 89 Solvency and technical reserve requirements. Regulations need to be in place for the industry to be able to take on and prudently manage additional risk. Reinsurance. Given that 98 percent of risks are currently rein- sured, the supervisors should make the oversight of the quality of reinsurance programs a high priority. The insurance supervisory unit has to build its credibility and staff capabilities. A major effort is needed to upgrade the institutional capacity for insurance supervision and the technical competence of the staff. In addition, it is critical to launch an education campaign to explain to the public the potential uses of insurance products. While the Insurance Association is best placed to take the lead in organizing such a campaign, the government should support its drive to do so where possible. Armenia currently has a pay-as-you-go, first pillar, defined benefit pension system. The system provides quite low pension benefits and is widely seen as unfair because it does not link the level of benefits to the level of wages (averaging a replacement value of just over 20 percent of average salaries). Pension benefits total only about 3 per- cent of GDP. The low pensions are due to (i) demographics, including an aging population and emigration by working-age Armenians; and (ii) tax evasion by employers and employees in the large informal economy. The contributor base is less than half of the working popula- tion and the ratio of contributors to pensioners is virtually 1:1. The government has formed a working group on pension reform, which is preparing a concept paper for transition to a three-pillar pension system and has already conducted actuarial projections on reform options. The concept under consideration includes replacing the current system with a compulsory funded pillar and establishing the framework for voluntary funded pension insurance while ensur- ing a minimum level of pensions for all through budget support. Con- tributions to the mandatory and voluntary pillars would be placed in individual accounts and invested in financial instruments. However, there are a significant number of considerations for the government to keep in mind in contemplating the direction, shape, and pace of pension reform, including the following: Conflicting objectives. The objectives of the plan involve signifi- cant tradeoffs among key policy objectives. In particular, during the transition period (which can last 30 years), there would be a direct tradeoff between utilizing reserves to increase current pensions or using them to fund individual accounts. Finding the appropriate balance between these objectives could be important for achieving public support for reform. 90 Part I: Policies to Sustain Economic Growth Fiscal policy sustainability. During the transition period to a compulsory funded pillar, supplementary budget support may be needed to pay current pensions and fund mandated levels of contributions to individual accounts. However, budget support should be provided only within a sound fiscal policy framework and appropriate limits on the size of public debt and debt servic- ing requirements. Lack of financial instruments in Armenia's financial sector. Armenia has few financial instruments that pension funds could invest in aside from bank deposits. The government debt market and corporate securities markets are quite small. In the initial years of funded pensions, a substantial portion of assets would have to be invested in foreign assets. The need to develop regulatory capacity and expertise. Funded pension systems are usually privately managed and have to be regulated. Before such a system could be established in Armenia, the government would need to form a regulatory body and build up its expertise. An appropriate financial institution with the requisite capacity would have to be found to act as custodian for pension fund assets (an Armenian bank that intended to receive deposits from pension funds should not be eligible). Scale of contributions. Given the small size of Armenia's pension system, the size of the pool of funds that would be created in the initial years of funded pensions would probably be too small to attract interest from qualified foreign pension fund managers, or to be financially viable in light of the fees that would be charged. A transitional set of arrangements might be needed during a start-up phase before a critical mass of funds was achieved and management by private pension companies was implemented. The transitional arrangements should also cover the creation of pension plans by interested employers. Public skepticism of financial institutions. While there is discon- tent with the current pension system, the Armenian public also holds some distrust of financial institutions as a consequence of previous experience with bank failures. There could be doubts on the part of the public that contributions to pension funds would ultimately be paid. A gradual transition could be benefi- cial, under which private accounts would accumulate slowly in the initial years, allowing confidence to grow over time. The Armenian housing finance market is at a promising early stage of development, but risks and weaknesses exist that could be reduced with limited government intervention to meet the needs of the broad population. Banks lent an estimated US$7.1 million in 2003--1.5 per- Finance as a Barrier to Accumulation 91 cent of total bank assets. Most mortgage lending is to upper-income households, in U.S. dollars, at fixed interest rates ranging between 14 and 18 percent, for low loan­to-value ratios of 50 percent, at an aver- age short-term maturity of five years. The main focus of reforms to housing finance should be on the development of the primary market for moderate- and low-income households. Key areas of reform include improving the technical capac- ities of lenders, enhancing public sector processes and systems, and improving the legal and regulatory framework. The government has proposed important reforms to strengthen the legal basis for property markets and mortgage lending, including defining real estate as a com- bination of land and structures, and improving the enforceability of the mortgage pledge. These reforms should be rapidly adopted. The government should shorten and simplify the process for title transfer and the registration of the mortgage pledge. The Central Bank should adjust provisioning rules to address the mismatch between the currency denomination of mortgage loans and the denomination of borrowers' incomes. The government and banks should more systematically gather and publicly disclose data on housing prices, transactions, mortgage lending, and mortgage performance to improve transparency and liquidity. The government should promulgate consumer disclosure reg- ulations on loan terms and risks, especially for non-dram mortgages. SUMMARY AND KEY RECOMMENDATIONS Financial intermediation is shallow and the cost of funds is high. Ratios of both deposits and loans to GDP are amongst the lowest in the CIS countries, and the interest rate spread is high. The cost of intermediation is raised by weak corporate governance, where even first-generation reforms related to transparency, disclosure of owner- ship, and accounting are still to be implemented. Creditor rights are being reformed but have been weak thus far, and the framework for secured transactions has been largely missing. Moreover, institutions to support the flow of information to support credit risk analysis need strengthening. Current reforms that are in progress, if implemented with determination, would lead over time to a lowering of costs and to a greater provision of bank credit to support growth. Implement a new law on corporate governance of banks. Strengthen official disclosures of direct and indirect ownership. Require by Armenian law the disclosure of beneficial owners. Ensure public access to the company registry and lists of found- ers and shareholders. 92 Part I: Policies to Sustain Economic Growth Implement a law on creditor rights and secured transactions with streamlining of judicial procedures. Strengthen the supervision of capital markets. Implement the unified supervision of financial and capital mar- kets, with new standards for the supervision of insurance, pen- sions, and housing finance institutions. Promote the development of a credit bureau. NOTES 1. Some advocate "relationship lending," in which banks make loans on the basis of the analysis of business plans and the borrower's history as an entrepreneur. This is unlikely to happen. First, most lending to businesses in the United States is secured by collateral. Second, the skills available for drawing up business plans are scarce. Third, those who have run businesses successfully are often the entrenched interests in Armenia, so that loans go to those who have tight control of the business sector. 2. Djankov, McLiesh, and Shliefer (2005), referred to in the preceding sec- tion of this chapter. Chapter 5 Impediments to International Integration I n chapter 3, the impact of the weak framework for competition on economic efficiency and growth was discussed. The deficien- cies in competition policy also affect the integration of the Armenian economy into international goods, services, and factor markets. For a small economy, the key to fast economic growth is integration into global markets. While Armenia has entered into a diamond global value chain (see Table 5.1), it has failed to exploit the opportunities offered by participation in other global networks, and other oppor- tunities for production and distribution, owing to weaknesses in the business environment. Given the country's geographic isolation, global integration depends vitally on export opportunities in the services trade, which has grown impressively. Armenian participation in the international outsourcing of ICT-related services as well as other services sectors (such as tourism and transportation) is below its potential owing to lack of competition and regulatory constraints. The composition of exports illustrates three structural weaknesses in international integration. First, exports of unskilled labor-intensive products appear low given Armenia's abundance in low-skilled labor. Moreover, other countries that are less intensive in low-skilled labor have a higher share of low-skill intensive exports than does Armenia. Thus, the share of unskilled labor-intensive products in Armenia's total exports was 1 percent in 2003, compared to 15 percent of Czech exports, 28 percent of Lithuanian exports, and 41 percent of Alba- nian exports. The deviation is smaller when set against the perform- ance of other CIS economies: Georgia has the same share as Armenia, Kazakhstan's is even lower (0.3 percent), and the Kyrgyz Republic's is 15 percent. This second group of countries (like Armenia) has yet to 93 94 Part I: Policies to Sustain Economic Growth Figure 5.1 Armenian Exports, 1997­2004 600 500 400 exports of other goods million 300 US$ 200 exports of diamonds 100 0 1997 1999 2000 2002 2003 2004 year Source: NSS. become part of the global garments value chain. Armenia has succeeded in entering the diamond value chain (characterized by skilled labor activity) but has failed to capitalize on its large reserves of cheap, unskilled labor. While garments have accounted for a huge share of manufactured exports from European transition economies, this has not been the case with Armenia. Second, Armenian firms have not become integrated into the pro- duction and distribution networks. Trade in such networks as fur- niture, the automotive industry, and the information revolution (or e-network) has driven export expansion for recent Central European entrants to the EU. Despite outstanding success in the development of the ICT sector, trade in the e-network (which includes electronics, office equipment, and telecommunications) has remained modest at best. This is also the case in the furniture sector, which is less tech- nologically demanding. Furthermore, not a single Armenian firm has become a supplier to a major producer of automobiles. Third, ICT firms have not been able to take full advantage of the recent change in the ICT sector worldwide. With the disappearance of the global "one-stop-shop" industry structures in the early 1990s,1 opportunities have emerged for new entrants. Armenian firms have Impediments to International Integration 95 successfully entered software and imaging technology niches. How- ever, they have failed to enter other stages of the production and delivery processes, such as providing front-end customer contact and support services or supplying components (such as metal, plastics, and electronics components).2 The cost of missed opportunities seems to be significant. The incor- poration of local producers into the clothing chain, production and marketing networks, and the supply of ICT consumer services usually brings new technologies and managerial know-how as well as direct access to larger markets, and thus benefits of economies of scale. Incor- poration boosts exports without local firms having to incur marketing expenses and provides for greater stability in earnings, thanks to the global reach of a "parent" company. The fragmentation of production eliminates the need to gain competency in all stages and aspects of pro- duction and allows a small country to focus on a subset of activities. At the same time, production sharing can broaden the range of final products whose components are produced in the small country and thus protect the country from a demand shock to a particular good. The reasons for failure to tap into these opportunities can be related to the unfinished agenda of demanding and complex structural reforms together with trade facilitation measures. This unfinished agenda con- stitutes a barrier to a deeper involvement of Armenian firms in a finer global division of labor based on just-in-time production, inventory management, and complex Internet-based communication links. The fundamental impediments to a deeper international integration are discussed in this chapter. CONTESTABILITY OF DOMESTIC MARKETS The major constraints to the contestability of Armenia's domestic markets for goods, services, and capital are lack of transparency and inconsistent implementation of regulations and laws.3 The legal arrangements concerning entry to these markets are liberal with one exception--telecommunications services are not open to external competition. Access to other services sectors is liberal and is locked in as a condition for accession to the WTO in 2003. Apart from the temporary exemption of telecommunications services to most-favored nation (MFN) status, other sectors have been open to most modes of supply, with the usual caveat concerning restrictions on the employ- ment of foreigners.4 No legal restrictions are in place on foreign capital inflows. The for- eign investment regime, governed by the 1994 Law on Foreign Invest- ment, provides for national treatment, MFN, and full repatriation of capital and earnings. The law banning foreigners from owning land 96 Part I: Policies to Sustain Economic Growth (they can only lease it) is not restrictive, as companies registered by foreigners in Armenia as Armenian business entities have the right to own land. The barriers in the opinion of international businesses relate to the inconsistent application of rules for taxes, customs (espe- cially valuation), and regulation, especially in the area of trade. While these may not be prohibitive for large firms with easy access to high levels of state administration, they create conditions of unfair competi- tion for medium-size businesses and other market entrants and add to uncertainty. Access to Armenian markets for goods is liberal in terms of offi- cial border and behind-the-border arrangements. Tariffs are low not only by CIS standards but also by international standards. Although Armenia's weighted and unweighted average MFN tariff rates are twice as high as those in the EU, they are still well below 5 percent (Table 5.1). Furthermore, commitments made upon accession to the WTO have infused a considerable degree of stability and predictability into Armenia's foreign trade policy. They have also reduced (but not com- pletely eliminated) the potential for capture of foreign trade decisions by narrow interest groups by providing the government with tools to tame rent-seekers. Armenia's two-band tariff regime, with applied MFN tariff rates at low levels of zero or 10 percent ad valorem, has Table 5.1 Average Applied and Bond MFN Tariff Rates in Selected Countries (in percent) European Kyrgyz Union Armenia Republic Lithuania (15) 2001 2002 2003 2002 Total goods Simple average (%) 3.3 4.5 1.3 1.5 Weighted average (%) 2.2 3.2 0.6 1.4 Agricultural goods Simple average (%) 8.3 7.1 4.7 3.2 Weighted average (%) 6.6 6.6 4.3 2.9 Industrial goods Simple average (%) 2.9 4.3 1.0 1.3 Weighted average (%) 1.1 2.7 0.3 1.3 Bound rate Simple average (%) 8.5 7.4 9.2 3.9 of all goods Weighted average (%) 9.6 6.4 9.4 3.0 Binding Coverage (%) 100.0 99.9 100.0 100.0 Memorandum: simple average tariff rate in 2004 3.0 5.2 1.5 1.5 Sources: Based on the UNCTAD TRAINS (http://www.unctad.org/Templates/Page. asp?intItemID=1907&lang=1), WTO IDB database (http://www.wto.org/english/ res_e/statis_e/its2005_e/its05_toc_e.htm), and the IMF trade information database for 2004 (http://www.imf.org/external/data.htm). Impediments to International Integration 97 been locked, thanks to accession to the WTO. Armenia's schedule of MFN "bound" tariff rates has seven ad valorem rates: 0, 4, 5, 6.5, 8, 10, and 15 percent.5 However, as a condition of accession, Armenia has been required to bind tariffs across all Harmonized System items. In addition, Armenia has committed itself, as a result of direct pressure from members of its WTO working party, to review periodically the specific tariff rates to assure that these do not exceed their equivalents of ad valorem bound tariff rates.6 Armenia has made significant commitments toward a liberal trade regime in the context of WTO membership, including commitments under the WTO General Agreement on Trade in Services (GATS). Except for telecommunications, Armenia's bound sectoral commit- ments are extensive in terms of both coverage and market access across different modes of supply of services. The number of subsectors in which exceptions are placed on a mode of supply (that is, unbound) is very small in Armenia's schedule. Except for settlement and clearing services for financial assets, including securities, derivative products, and other negotiable instruments, there are no exceptions under deliv- ery mode 1 (cross-border). Consumption abroad (mode 2) is subject to no restrictions. The only restriction in market access under mode 3 (commercial presence) is the requirement that suppliers of technical testing and analysis services should be legal entities constituted under Armenian legislation. In contrast to a number of countries--including highly developed economies--that limit a maximum foreign equity stake, there is no such a limit in Armenia. Given the emphasis in national economic policy on promoting the ICT sector, and given Armenia's specialization in computer services that sets it aside from countries at a similar level of economic devel- opment, it is puzzling that Armenia did not join the Information Technology Agreement (ITA) upon accession to the WTO. The ITA, concluded by 29 participants at the Singapore Ministerial Conference in December 1996, provided for participants to completely eliminate duties on ITC products covered by the Agreement by January 1, 2000. The number of WTO members that participate in ITA has risen to 63 states, including China, India, and East Asian countries includ- ing Malaysia, Singapore, and Thailand. Participation does not assure knowledge-intensive growth but can help to create the right condi- tions for such growth. However, neither commitments under the WTO Agreements nor legal provisions protecting private property rights and enforcing con- tracts alone have ensured the high contestability of Armenia's domes- tic markets. First, the capacity of courts remains weak, as they operate slowly in the enforcement of contracts and the mediation of commer- cial disputes. Furthermore, firms do not trust the ability of courts to 98 Part I: Policies to Sustain Economic Growth act independently and enforce decisions. Neither do they trust the impartiality of the state administration, regarded by businesses as cor- rupt. Second, the computerization of customs has to date fallen well short of improving the quality of customs services, with customs pro- cedures yet to achieve the WTO standards of transparency. The lack of the bilateral harmonization of customs practices with Georgia, cor- ruption, and the limited application of ITC further exacerbate trans- action costs. Nontransparent procedures corroborate businesses' complaints about the continuing practice of customs of extracting "irregular" pay- ments from them. Corruption and the imposition of unofficial fees at the border are most frequently reported by the private sector as serious issues. The fact that customs clearance procedures are applied with equal zeal also to exports is suggestive, as, generally, in other countries, customs efforts are largely directed only at imports. For example, on the rail system from Armenia to Georgia, unofficial fees account for approximately 6­13 percent of the total costs of transport (Molnar and Ojala 2003). A typical container shipment by truck from Tbilisi to Rotterdam is subject to unofficial payment costs totaling 7 to 40 percent of the total logistics costs, with customs clearance being the most significant element. The administration of VAT rebates to exporters is an important fac- tor that deters exports. VAT reimbursements do not occur quickly enough and are underpaid by the government. A recent survey (World Bank 2004a) notes that only 44 percent of surveyed firms entitled to the refund have claimed it.7 One obvious reason for unclaimed refunds is that for the remaining 56 percent of companies it took, on average, 145 days to receive a VAT refund. Moreover, on average they received 15 percent less than the amount to which they were entitled by the law. In 2005 a significant improvement has occurred in the pace and quantity of refunds. Moreover, the most anticipated benefits of the computerization of customs services (that is, reducing the release time of consignments and slashing the documentation or bureaucratic burden put on a trader or customs broker) have so far failed to materialize. Traders and customs brokers have limited access to the system. The old prac- tice of bureaucratic delays fueling corruption, combined with a lack of capacity in customs administration, continues. Other barriers to the high contestability of Armenia's domestic markets stem from the high "hassle" cost of doing business in Arme- nia discussed earlier. Weaknesses in tax administration, the capacity of the CPEC to enforce competition laws, and the weak judicial system have been responsible for low FDI inflows and for Armenia's limited success in attracting activities outsourced by multinational corpora- tions (except for diamonds). Impediments to International Integration 99 REFORMING CUSTOMS RULES Ideally, the government should extend the same customs rules that it applies to diamond cutting to other sectors of the economy. This would entail reducing the time for customs clearance to one day, accepting the invoice value, and conducting in-house clearance, if needed. While these steps would be a dramatic improvement over the current practice, other transaction cost­raising elements would have to be addressed as well, such as the VAT refund system mentioned previously. Extending a "diamond-like" regime with an improved VAT refund mechanism to all exports would be difficult. While inputs crossing customs borders are usually subject to tariff rates and indirect taxes if sold domestically, the tariff rate on diamonds is zero and no taxes are levied. By the same token, a domestic firm has no incentive to sell the imported input and will domestically circumvent customs, thereby depriving the budget of customs and other tax revenues. In short, there is no need for duty drawback, temporary admissions, or bonded warehouse schemes, which require considerable capacity on the part of customs administration to be really effective. While compu- terization--if accompanied by administrative reforms and changes in customs procedures commensurate with the newly acquired techno- logical capacities--may significantly improve this capacity, this might take some time. Therefore, the suggested reform would be to move gradually and establish a diamond-like regime in terms of administrative efficiency for a selected group of firms. The selected firms would be put on a "white list." Firms on such a white list would be eligible for special treatment by the customs and VAT administration. Customs would be obliged to complete customs clearance within a very short period; it would also have to observe the provisions of the WTO Agreement on Customs Valuation instead of using reference prices; and it would be obliged to run an effective scheme that provides duty waivers and exemptions from other restrictions on imported inputs for estab- lished exporters that import inputs, whether for domestic production or export, or both.8 The VAT administration would be obliged to rebate the VAT to a white-listed firm on the same day that exports are cleared by customs. For any delays, a firm would be entitled to a refund plus interest. Which firms, in addition to diamond-cutting firms, should be put on a white list? While the criteria to be used may be subject to fur- ther refinement, the general guidelines are easily discernible. First, the process and the criteria should be transparent and open to the public scrutiny of press and nongovernment organizations, such as business associations. 100 Part I: Policies to Sustain Economic Growth Second, foreign-owned firms publicly traded on the U.S. and EU stock exchanges should be automatically included. Their operations are subject to the highest standards of international scrutiny and they can ill afford to enter into shady business operations. This would take care of participation in the global networks of production and distri- bution, simply because their major movers are publicly traded, large, multinational corporations. Third, firms operating in the ICT sector should also be included in the white list as their use of imported inputs is limited and the value added created locally is huge. Their exports are not subject to customs clearing procedures. In fact, they are virtually impossible to monitor. However, in order to claim duty and VAT rebates, they would have to disclose some transactions to the financial authorities. The overall direct benefits to ICT firms of being included in the white list are lower than for exporters of manufactures, although firms of this sector are as vociferous in their criticism of the VAT rebate scheme as are the diamond-cutting firms. Furthermore, this would be a significant step in improving the business climate for a sector critical to Armenia's move to a knowledge-intensive economy. Fourth, for the remaining firms, one would have to identify one or two criteria related to a record of past dealings with the tax adminis- tration. One criterion might involve the number of years in existence; another might take into account the past record of dealing with tax and customs administrations. Both would have to be clearly defined. The general idea would be to reward law-abiding firms. Firms that do meet the criterion of the length of existence but are involved in inward processing (for example, of garments or footwear) might be exempted. Representatives from government and other relevant business administration would annually or semiannually update the white list. A mechanism to address the grievances of those that have not been included should be designed and implemented. However, if the crite- ria are precisely defined and transparently implemented, the number of cases handled under this mechanism would be rather limited. A white list alone would not be sufficient for improving customs efficiency. There are other areas which might be relatively quickly fixed. One quick action would be the establishment of an independent professional association of customs brokers. The association would be empowered to license and scrutinize its members. In order to assure integrity and responsibility, it would operate according to a transpar- ent code of conduct that would be obligatory for all members and it would have an institutional voice in the government bodies respon- sible for overseeing customs administration. Another quick action would be to open the customs computer sys- tem to customs brokers. Customs-related documents would be brought Impediments to International Integration 101 in line with what is really required under a computerized system such as the Automated System for Customs Data (ASYCUDA). A further action would be the simplification of customs clearance procedures. Under normal circumstances, the customs clearance of exports serves one major purpose: to assure that the shipment does not contain prod- ucts banned for exports. Once an exporter produces a certificate of origin issued by the Chamber of Commerce, the shipment should be immediately cleared. Hence, a strict time limit on releasing a shipment should be introduced; if the time limit is exceeded, a shipment should be immediately released. The above measures are relatively easy to implement. While they require the close attention of policy makers, they consume adminis- trative resources in a very moderate way. Over the long term, they would actually reduce the administrative burden. INTEGRATION IS BURDENED BY EXPENSIVE BACKBONE SERVICES The domestic elements of the business climate (that is, regulatory regimes and measures directly affecting entry, conduct, and exit for firms) also influence the external interaction of domestic firms. Ele- ments that are as important for domestic as for external activities include procedures for the simplification and harmonization of inter- national trade; the state of the infrastructure and its management; and the provision of such backbone services as telecommunications, bank- ing, insurance, transportation, and business services. Other elements are customs procedures, related regulatory procedures for border clear- ance, technical standards regulations, and airport efficiency and effec- tiveness. These elements shape the ease and speed with which goods and services move across national borders services; therefore, they are crucial to trade in goods. Consequently, improvements in the domestic business climate may produce limited positive economic effects if not accompanied by similar improvements in trade facilitation. Weaknesses in the provision of backbone services add greatly to the costs for domestic firms to participate in the emerging division of labor based on international outsourcing, just-in-time production, and supply-chain management. Increasingly, sliced value chains, with individual production stages being moved to countries with compara- tive advantages, have become trademarks of a current global economic landscape. Interaction among "production blocs" of border-spanning production networks is particularly vulnerable to delays in and disrup- tions between individual stages of the supply chain owing to weak- nesses in service links. Hence, the poor quality of backbone services and trade facilitation deters foreign firms from incorporating domestic 102 Part I: Policies to Sustain Economic Growth firms into their supply chains and also acts as a barrier to other types of trade. In a nutshell, these are the factors responsible for the emer- gence of trade within global value chains and networks. The factors responsible for the emergence of a new form of the divi- sion of labor driven by the globalization of production appear to be missing in Armenia. Falling transportation and communications costs have created opportunities for the outsourcing, just-in-time production, and supply-chain management that have been altering the competitive landscape of many countries by relocating business activities and pro- viding a new source of entry into international markets. However, high transportation and communications costs as well as high transaction costs of doing business in Armenia are barriers to participation in the division of labor based on production fragmentation. The limited presence in the global value chain and trade network of garments, and the almost complete absence of Armenian providers for some ICT customer support services, have not been the result of adverse external conditions. These conditions are mostly home made, and thus can be addressed by changes in government policies. While geographic or political constraints are difficult to overcome, contempo- rary technology combined with a right mix of policies may consider- ably ease their negative impact. Outsourcing, just-in-time production, and supply-chain management, which are all critical to transferring abroad a slice of a production process, cannot function efficiently if there are delays and disruptions because of weaknesses in the service chain. The weakest link in the service chain in a country may tip the balance against including a firm in the global supply chain. The serv- ice chain includes backbone services, namely, telecommunications, transport, financial services (banking, insurance, securities trading), distribution and business services (legal services, accounting, consult- ing), as well as customs procedures. Neither politics nor geography has affected Armenia's strong and expanding presence in the diamond value chain. This is the result of inherently low transport costs and friendly government policies that are not extended to other sectors. Potential or actual producers of parts and products for other chains and networks are not in a similarly privi- leged position. Transportation costs usually account for a large share of total production costs. Few items match diamonds or services' com- bination of a lightweight and high-value unit. Just-in-time production and supply-chain management are less important in the diamond busi- ness, as is the timeliness of the communications links. For products other than diamonds the distance from major markets clearly matters, as trade costs, including transportation, customs, and communications, may make operations noncompetitive in world mar- kets. But even though trade has expanded more rapidly in areas geo- Impediments to International Integration 103 graphically closer to the major world markets, trade in more distant regions has also grown, which demonstrates that policies that reduce transaction costs can soften the negative impact of distance. Air and telecommunications links play a crucial role in this reduction. Wilson, Mann, and Otsuki (2004) identify four indicators of the capacity to facilitate trade. These are port efficiency, the customs envi- ronment, the regulatory environment, and the ICT infrastructure. Port efficiency measures the quality of the infrastructure of maritime ports and airports. The customs environment measures the direct customs costs as well as the administrative transparency of customs and bor- der crossings. The regulatory environment measures the country's approach to regulations and their quality. The ICT infrastructure meas- ures the extent to which an economy has the necessary domestic infra- structure (such as telecommunications, transportation, financial inter- mediaries, and logistics firms) and is using networked information to improve efficiency and to transform activities to enhance economic activity. Since we have discussed regulatory environment and customs practices, and since a private operator runs the airport in Yerevan, we shall focus on ICT infrastructure and transportation services. ICT INFRASTRUCTURE: THE PRIVATE TELEPHONE MONOPOLY Rapid technological development in the electronics, computer, and telecommunications industries has eroded the previously inherent natural monopoly characteristics of telecommunications. By the same token, it has weakened somewhat the negative impacts of the govern- ment's decision to grant a legal monopoly (originally until 2013) to ArmenTel, the local telecommunications company, owned by the Hel- lenic Telecommunications Organization. The monopoly also encom- passes cellular; the local loop providing land line links to firms and households; and international sources, including, theoretically, Internet services. The only areas that have escaped the monopoly's reach, albeit not completely, are the Voice over Internet protocol (VoIP) and private international satellite communication such as that operated by Lycos, a private foreign company. These developments account for the low volume of officially captured international calls. But satellite communi- cation is not a full substitute for access through land lines, and satellite has nontrivial fixed costs and cannot be connected to the telecommu- nications network to provide backup and higher utilization. Therefore, while modern technologies offer some ways of circumventing the monopoly reach of ArmenTel, it controls the decisive "last mile" local loop access and has dominated the mobile phone market. 104 Part I: Policies to Sustain Economic Growth How has ArmenTel's monopoly affected the development of tel- ecommunications in Armenia? One way of addressing this question is to compare the costs and the use of telecommunications services in Armenia to those in other CIS countries. This may not be a very demanding benchmark, as many other CIS countries have equally poor regulatory reform and privatization and some still maintain a state monopoly for service provision. Nevertheless, the general con- clusion is that, overall, Armenia's telecommunications services com- pared with other CIS countries performed poorly over 1998­2002. Despite an increase in 2002 over 2001, in 2003 the international outgo- ing traffic from Armenia declined significantly more on average than in other CIS countries. The cost of using the Internet in Armenia was 41 percent higher than the average for the CIS. The prices for high- speed connections charged by ArmenTel are around 30 times more than those in countries where telecommunications services are not monopolized. These developments explain why Armenia had a low number of Internet users, well below the CIS average. Moreover, the number of cellular phones per 1,000 people is significantly lower than in neighboring Georgia. Tables 5.2 and 5.3 compare the telecommunications situation in Armenia with those in Azerbaijan and Georgia. Tables 5.4 and 5.5 compare benchmarks for telecommunications infrastructure and the Internet in Armenia with selected countries. The poor quality of the ICT infrastructure is a barrier to the devel- opment of the economy and to its shift toward the higher knowledge intensity and enhanced networking that technological change permits Table 5.2 Telecommunications Sector Overview Armenia Azerbaijan Georgia Population, 2003, millions 3.80 8.23 4.89 GDP per capita 2003, PPP, US$ 3,607 3,606 2,569 Telecom revenue, 2002, US$ millions 70.1 85.8 135.0 Telecom investments, 2002, US$ millions 22.7 28.7 n.a. Number of fixed lines, 2003 563,679 941,366 650,500 State ownership share in the incumbent ArmenTel 10% n.a. n.a. fixed-line operator, September 2004 (90% owned by Greece's OTE) Number of mobile subscribers, June 2004 137,530 1,317,560 779,300 Number of mobile operators, June 2004 1 2 3 Sources: World Development Indicators (WDI) (http://web.worldbank.org/WBSITE/ EXTERNAL/DATASTATISTICS/ 0,,contentMDK:20398986~menuPK:64133163 ~pagePK:64133150~piPK:64133175~theSitePK:239419,00.html), ITU, European Mobile Communications Report (http://www.mac.doc.gov/ceebic/countryr/Fyrm/ MARKET/ Macedonia%27s%20Information%20Technology%20Sector.pdf). Table 5.3 Percentage of Households with Access to a Telephone Income quintiles Country Total 1 2 3 4 5 Capital (Yerevan) 83.7 73.6 80.6 78.9 88.4 92.8 Armenia Other urban 64.8 46.8 59.5 67.5 76.8 79.9 Impediments (ISLS 2001) Rural 44.0 34.8 40.4 42.8 48.1 53.3 Whole country 61.8 50.4 56.4 59.4 68.1 74.5 Capital (Baku) 63.4 58.4 55.1 62.8 68.5 71.6 Azerbaijan Other urban 44.4 43.2 42.7 40.8 48.2 48.7 to (HBS 2001) Rural 18.4 18.2 17.3 18.3 17.9 20.1 International Whole country 38.6 38.4 36.9 35.8 38.9 43.1 Capital (Tbilisi) 71.7 59.1 63.7 70.9 76.9 82.1 Georgia Other urban 47.3 31.8 44.1 49.9 58.6 66.7 (HBS 2001) Rural 8.8 5.0 8.2 8.8 10.5 10.8 Integration Whole country 34.7 25.4 33.2 34.5 37.9 42.5 Source: Integrated Survey of Living Standards (ISLS); Household Budget Survey (HBS). Note: Comparisons should be made only within countries and not between them, since the surveys were conducted independently and the questions were not identical. 105 106 Part I: Policies to Sustain Economic Growth Table 5.4 Basic Telecommunications Infrastructure Benchmarks GDP per Households Mobile capita with a Main lines penetration, 2003, television, per 100 August PPP 2002 inhabitants, 2004 Country (US$) (%) 2003 (%) Albania 4,571 90 8 32 Armenia 3,607 91 15 4 Azerbaijan 3,606 121 11 17 Bosnia and Herzegovina 6,029 87 24 31 Bulgaria 7,807 93a 38 59 Croatia 11,139 94d 42b 60 Czech Republic 16,448 99 36 100 FYR Macedonia 6,762 82d 27b 40 Georgia 2,569 76 13 17 Greece 19,973 98c 45 96 Hungary 14,572 96c 35 76 Moldova 1,505 -- 22 14 Romania 7,222 97 20 40 Serbia and Montenegro -- 92 24 37 Slovak Republic 13,469 100 24 75 Slovenia 19,300 91 41 95 Turkey 6,749 108 27 45 Sources: WDI, ITU, European Mobile Communications Report, Paul Budde Communication. a. Data from 2003; b. data from 2002; c. data from 2001; d. data from 2000; -- =not available. and encourages. The absence of high-quality services and the high costs of existing services affect all sectors of the economy, amounting often to the equivalent of a prohibitive tax on some potential markets. It is impossible to estimate how many transactions have not taken place because of this situation. Neither is it possible to tell how many industrial or service operations would have been transferred to Arme- nia had there been more efficient telecommunications services. It is, however, indisputable that very high telecommunications costs have severely exacerbated the disadvantage associated with Armenia's geo- graphical location. High-quality ICT infrastructure and services are crucial to the development of a contemporary economy. Impediments to International Integration 107 Table 5.5 Internet Benchmarks Internet International Internet PCs users Internet hosts per 100 per 100 bandwidth per 1000 inhabitants, inhabitants, per capita, inhabitants, Country 2002 2003 2002, bps 2003 Albania 1.2 1 3.9 0.08 Armenia 1.6 4 2.1 0.55 Azerbaijan 4a 0.3 0.07 Bosnia and Herzegovina n.a. 3a 6.6c 1.89 Bulgaria 5.2 21 10.1 6.66 Croatia 17.4 23 41.2 6.78 Czech Republic 17.7 31 2,189.1 27.44 FYR Macedonia 4.5 5a 24.2 1.73 Georgia 3.2 3 1.01 Greece 8.2 15 222.0 17.05 Hungary 10.8 23 246.3 35.78 Moldova 2.1 8 9.4 3.32 Romania 8.3 18 89.3 2.18 Serbia and Montenegro 2.7 8 0.9b 1.84 Slovak Republic 18.0 26 1,516.0 21.22 Slovenia 30.1 38a 539.6 21.48 Turkey 4.3 8 16.3 5.08 Source: ITU. Online at: http://www.mac.doc.gov/ceebic/countryr/Fyrm/MARKET/ Macedonia%27s%20Information%20Technology%20Sector.pdf a. Data from 2002; b. data from 2001; c. data from 2000; bps = bits per second; -- = not available. The ArmenTel monopoly has been reduced and further cellular competition will be allowed in 2009. A new mobile operator has been licensed, although, given the objectives pursued and the process fol- lowed, it is unlikely that a competitive regime will emerge. Duopolies can produce outcomes close to monopolies unless competition law is actively applied to prevent tacit collusion. The PSRC has been given enhanced powers in an attempt to countervail the adverse effects of ArmenTel's continuing dominant market position. There is little expec- tation that Armenia will benefit fully from the rapid fall in costs and the introduction of new services that a fully liberalized market would offer. It is understood that the government regards current policies as the best achievable until the ArmenTel license expires in 2013. It is 108 Part I: Policies to Sustain Economic Growth crucial that every opportunity be taken to maximize the possibilities for competition given the current policy and agreements. Clearly, the government faces the difficult issue of minimizing the negative impact of some of the most restrictive provisions of its con- tract with ArmenTel without endangering the credibility of Armenia's commitment to the sanctity of contracts or facing a considerable fiscal cost. The policy so far has been a combination of a strong commit- ment to respect the contract while requiring that ArmenTel meets its obligations, together with a declaration that regardless of its outcome, mobile (cellular) and data services sectors will be open to domestic and foreign companies in the future. In the meantime, the tactic is being followed of increasing the regu- latory pressure on ArmenTel by placing telecommunications under the jurisdiction of the PSRC, which deals also with power, water, and natural gas. However, as long as ArmenTel essentially has monopoly rights, the major task would be overseeing a monopoly and trying to use regulatory tools to encourage the utility to expand its out- put and lower its prices to efficient levels. The management of the license should ensure that all provisions concerning the commitments of ArmenTel are implemented. This task is different from setting the regulatory framework that facilitates entry and protecting the com- petitive process. Hence, as long as the monopoly is in place, the best intervention is to ensure that the PSRC is an efficient and effective regulator and enforcer of the modified agreement with ArmenTel. With the current duopoly likely to continue until 2009, an additional option would be for the government to develop and announce now its future policy: namely, that all government-imposed economic entry barriers would be removed when the license expires (so that, for example, the electric- ity distribution company could provide telecommunications services if technically and commercially viable). Furthermore, the available and necessary radio frequency spectrum will be auctioned off well before 2009, so that further mobile operators, to the extent that they consider they would be commercially viable, would be able to enter the market as soon as restrictions are lifted. Private networks (includ- ing those using satellite links) could prepare to interconnect and pro- vide services to other customers. The market after 2013 should be regulated only by competition law. AIR TRANSPORT: AN UNTAPPED ASSET Despite the importance of land transport, the infrastructure system in Armenia requires modernization, including adequate provisions for ongoing maintenance and management and considerable road reha- Impediments to International Integration 109 bilitation. In addition, most of its rail track and rolling stock are in need of repair or replacement. Since rail and road links from Georgia to Armenia account for 70 percent of Armenian trade, their quality is of particular importance. They are all in poor condition. The lack of direct sea access, the mismanaged state-owned railway company (Holden and Sahakyan 2004), the high freight rates, and the ongoing blockade of the borders with Azerbaijan and Turkey, all contribute heavily to the high transportation costs faced by importers and export- ers, which puts Armenian firms at a competitive disadvantage. Unless transport costs go down, the trade patterns will remain biased toward goods with high value relative to weight at the expense of the bulky low-cost products of light industry. This may explain why, unlike the situation in other transition economies, Armenia's exports of textiles, apparel, and footwear have been limited. What can the government do about transportation costs? It is important to ensure good road operation and development in order to lower vehicle operating costs and transport time, and the govern- ment may be spending too little by way of public investment in this area (note that the medium term expenditure framework for 2006­8 suggests that funding for state roads is declining as a share of GDP). Placing the state-run railway on a business footing would allow it to compete more effectively with road transport and to maximize its economic value added. The government can influence the perform- ance of infrastructure service providers through changes in incen- tives and the institutional environment. Investments in infrastructure and strengthening the regulatory structure can reduce transport costs with particularly large potential for reducing costs in air transport and railways. Overall, however, as mentioned earlier, transportation costs, both overland and by air, are well above the levels in other CIS land- locked countries, including those located further away than Armenia from their most important markets, largely in the EU. The shift away from air transport to land transport indicates that the former has become less competitive. Providers of air cargo serv- ices have been unable to retain their clients or to lure new ones with more attractive prices, as freight shipped by air from Yerevan fell between 1997 and 2003 by more than two-thirds (depending on the indicator).9 The fall has been even more dramatic considering that the value of exports of goods rose 3.3 times over this period. Volumes increased significantly in 2004, back to the level four years earlier. The re-opening of the land corridor through Georgia might have boosted land transport. But had there been more aircraft belly space available, thanks to an increase in the number of passenger flights, lower air cargo rates would have attracted a much larger portion of shipments than was the case. This clearly has not happened, as there was no significant increase in the frequency of flights until 2004. It remains 110 Part I: Policies to Sustain Economic Growth to be seen whether the resumption of growth in air services contin- ues, is sustained, and translates into better services for cargo shippers (including lower prices). The fall in the share of the most dynamic worldwide mode of transport--air transport--can be attributed to the fall in the competi- tiveness of air transport services owing to poor infrastructure and a restrictive aviation policy. The poor infrastructure component has been potentially addressed, thanks to the government's decision to sign a 30-year concession agreement with Argentina-based Corpora- cion America for the management of the Yerevan Zvartnot Interna- tional Airport in 2003. Careful management of the concession by the government, within a clear policy framework, should result in a more efficient and effective provision of airport services. The removal of the infrastructure barrier alone will not suffice to make air transport competitive. The appropriate regulatory environ- ment must be in place in order to take advantage of better infra- structure. However, recent government action is not encouraging: an agreement has been signed with the Russian airline Siberia for the establishment of a new airline (Armavia) to take over Armenian inter- national air "rights" on an exclusive basis until 2013 (with a stream of annual payments to the government for these rights).This agreement suggests that the approach to aviation policy has remained restric- tive via exclusive and restrictive route rights. Instead of moving to a less restrictive approach that emphasizes the removal of govern- ment-imposed entry barriers to air services (as pioneered by Chile and implemented by countries such as Latvia), Armenia has retained the bilateral system of air transport regulations, which is based on a positive list approach that limits the provision of services to those that are explicitly permitted.10 An existing Armenian airline, the Armenian International Airline, has been forced out of serving the Armenian air markets as a result of the government's agreement with Armavia. Leaving aside safety and aviation security issues, which need to be addressed in any future scenario, the regulatory philosophy underly- ing the restrictive approach to bilateral aviation agreements has been to protect the "national" carrier from external competition--Armavia in this case. Governments have understandably been concerned about service continuity and have often considered that only a national car- rier will ensure such continuity. Costly regulatory assistance is usually given, as in this case, although the government, unusually by interna- tional practices, is receiving explicit annual payments in return for the grant of exclusive rights. Inferior services in terms of costs and qual- ity often result. International experience shows that in a deregulated environment airlines will serve markets: concerns about service conti- nuity have been generally unfounded. Regulatory restrictions impede this process and tend to exclude the most efficient airlines. Impediments to International Integration 111 The empirical evidence from countries that have deregulated the domestic aviation sector (such as the United States and the EU) is robust: passengers and air freight shippers in both the EU and the United States have experienced a dramatic and continuing decline in airfares. Thanks to the opening of the sector to new entrants, existing carriers have come under strong competitive pressures. These in turn have reduced costs through gains in productivity. The entry of the so-called low-cost carriers (LCCs) has halved airfares and stimulated volume by at least 70 percent in both the EU and the United States. Competition has also been responsible for a faster adaptation and diffusion of new technologies. The combination of liberalization and technological progress has been behind a 3.5 percent average annual decline in real terms in prices of services over 1991­2001. The costs of bilateral aviation agreements are not just reflected in the higher prices of air transport due to the absence or limited competition. In fact, they also involve the much more important and more difficult- to-estimate costs of forgone opportunities. These costs are potentially large as lower airfares boost tourism, stimulate important flows of ideas and human capital, deepen networks, and create new opportuni- ties for firms to market their products domestically and internationally. Tourists in turn increase the demand for a range of services and goods. The empirical evidence can be supplemented by several examples. The liberalization of aviation policy, together with investment in the Emirates airline, has contributed to Dubai's impressive economic development, which is driven largely by tourism and services. The result of the relatively liberal policies pursued by the United Arab Emirates since the 1980s is that Dubai is now served by around 100 airlines flying to 145 destinations and Emirates has become a formi- dable world-class competitor. More important, Dubai it has become a tourist and business center in the Middle East. Last, but not least, one suspects that the probability for a firm to become incorporated in global production and distribution networks increasingly on the basis of just-in-time production is low in the absence of reliable, high- frequency, competitive air transport services. A second example suggests that airline liberalization has been a contributing factor to economic development in Ireland--often described as a Celtic Tiger because of its spectacular economic growth performance over the last decade or so. It is often overlooked that the measures that created a "virtuous circle" leading to the Irish economic boom included not only public expenditure cutbacks that allowed for tax reductions but also EU-driven airline deregulation that allowed Ryanair to fully develop the LCC business model following the U.S. Southwest Airlines example. Deregulation "facilitated a more than doubling of tourist numbers over the following decade" and contrib- uted significantly to FDI inflows (Barry 2003, p. 909). 112 Part I: Policies to Sustain Economic Growth For a small landlocked economy aiming to maximize its long-term economic growth rate, the best aviation policy would be along the lines of the Chilean "Open Skies" policy, which would include Armenian carriers having reciprocal rights to compete. Under this arrangement the government policy would be to eliminate government-imposed entry barriers to air transport. This should be the option that Arme- nia should adopt as quickly as possible, although the agreement of Armavia would probably be required, unless their contract with the government is renegotiated. Full implementation of such a policy is not likely in the immedi- ate future, as it would take both governments (Armenia and Russia) to agree to liberalize a bilateral agreement. Taking into account the lack of Armavia's business interests in some routes and interests of other governments to open access for their carriers, Armenia may be able to strike Open Skies agreements with the United States, the United Arab Emirates, and Bahrain, for example. Since all three are potentially important markets, and since the latter two offer relatively close outlets to a wide range of markets, these agreements would be of significant economic importance to Armenia. The reason why the United Arab Emirates and Bahrain might be interested in Open Skies agreements is that both have carriers with well-connected hubs that offer high-frequency connections throughout the Middle East and Asia. Such an agreement would also offer an extra air cargo link to European markets. The reasons why the United States might accept the offer are two- fold: one related to the Armenian diaspora and commercial relations in the United States and the other to the international aviation policy of the United States. The pressure from the diaspora might be rein- forced by U.S. carriers interested in flights to Armenia. As for the U.S. aviation policy aspect, such a reciprocal agreement would be natural, as at present, the United States has Open Skies agreements with over 70 countries, including Uzbekistan and, recently, India. The U.S. type of Open Skies agreement typically includes the fifth freedom11 code- share rights for U.S. carriers extended beyond hubs outside the United States. This would allow a U.S. carrier to add its code to any partner airline (for example, an EU carrier flight into Yerevan). To date, the EU has displayed little interest in entering into negotia- tions to achieve the major modifications that Open Skies agreements would entail with small countries with restrictive air service policies. Rather, it has been seeking to extend arrangements like the European Civil Aviation Area to other countries with the Open Sky type of poli- cies (the United States, Singapore, and New Zealand, for example). Within the parameters of this approach, the CIS might be a potential partner but not Armenia alone. Impediments to International Integration 113 Table 5.6 Phased Civil Aviation Reform Time framework Incumbent carriers New entrant carriers April 1, 2006, to April 1, 2007 Up to 3 daily flights Up to a daily freq. From April 1, 2007 Unrestricted Unrestricted Source: Bank staff. Under these circumstances, the Open Skies agreement with the EU is currently an unlikely option, and Armenia should instead announce its objective of removing restrictions on carriers and inbound flights, subject to receiving reciprocal rights. If Armenia considers that the protection this would accord Armavia would be insufficient (includ- ing taking account of the agreement with Armavia), then depending on the assessment of the agreement with Armavia (or its successor), the government could take a phased approach toward a fully liberal and unrestrictive policy. The steps could be as shown in Table 5.6. The government could also implement a liberal approach for fifth freedom rights as well as explicitly announcing that it favors a multi- airline policy. Armenia could gain considerably from Open Skies arrangements among CIS countries modeled on the EU European Civil Aviation Area. But this would take time, as Armenia would have to develop the model and convince other CIS members, perhaps building a coalition with like-minded countries. A likely increase in the number of flights connecting Yerevan with the Middle East, the United States, and the EU would exert competitive pressure on airlines that service routes to Moscow and other destina- tions in Russia, particularly as the recent survey of passengers through Yerevan airport has shown that the great bulk of passengers flying to Moscow were actually en route to Europe and the United States. SUMMARY AND KEY RECOMMENDATIONS Armenia enjoys an admirably open trade and investment regime, but firms are insufficiently linked to the international production and dis- tribution chains, and ICT firms have not been able to make good use of international opportunities. Major constraints lie in the lack of transparency and consistency in regulation and the poor customs administration. Customs rules can be reformed to make its adminis- tration more rules-based. Policy-induced high costs in telecommuni- cations and in air transport need to be fought, as current structures impose a high "tax" on competitiveness. It is suggested that the 114 Part I: Policies to Sustain Economic Growth future liberal telecommunications regulatory regime be decided and announced now, and that rapid steps be taken to remove economic restrictions on civil aviation with the aim of developing an Open Skies policy in conjunction with the EU. Customs and VAT: Establish a "white list" of firms subject to spe- cial treatment by Customs and VAT administration, including quick customs clearance; use provisions of the WTO Agreement on Customs Valuation instead of using reference prices; create an effective scheme that provides duty waivers and exemptions from other restrictions on imported inputs; and give rebates of VAT as soon as exports are cleared by customs. Extend direct transfer input to all customs houses. Bring customs-related documents in line with what is really required under a computerized system of the ASYCUDA type. Simplify customs clearance procedures for exports with a strict time limit on releasing a shipment; if the time limit is exceeded, the shipment should be immediately released. Support participation in international R&D networks (such as the EU 6th Framework Program). Sign the WTO Information Technology Agreement. Publish a statement of government civil aviation policy for con- sultation. Introduce new and increased services in 2005­6. Pass a new telecommunications law and see that the responsibil- ity for implementation, including that of the ArmenTel license, is passed to PSRC. Have a cellular competitor enter service. Have the PSRC prepare a regulatory policy statement in 2006 that, among other things, lays out an implementation plan for the entry of a third service provider by 2009. NOTES 1. A large number of firms, connected through complex and borderless supply chains made up of high-volume, multicustomer, and multinational specialists at each level, have replaced the vertically integrated firms domi- nant in the 1980s and early 1990s. 2. The front-end, customer-centric portion of the supply chain includes service providers such as contact centers, order processing, and technical sup- port. At the same level is the order fulfillment hub, which is a combination of a logistics center with order picking and configuration capabilities. Impediments to International Integration 115 3. The following sources have been used in this section: the World Bank's various country reports, the World Bank Investment Climate Survey, BEEPS and BEEPS II surveys, Economist Intelligence Unit information, the Heritage Foundation's Index of Economic Freedom, and the World Bank's Doing Busi- ness Indicators. 4. Armenia, however, has made a commitment "to provide unlimited mar- ket access for all basic and value-added telecommunications services sub- sector initially covered by the monopoly immediately upon suspension or termination of monopoly rights in that subsector on or at anytime prior to the end of the monopoly." See WTO 2002. 5. The simple average bound tariff rate for all products is 8.6 percent, with the average rate of 14.9 percent for agricultural products and 7.7 percent for industrial products (authors' calculation based on Armenia's "Bound" Sched- ule). Both averages put Armenia in the middle of transition economies. For agricultural products, Albania, the Slovak Republic, and the Czech Republic (prior to the EU accession) have lower rates of 9.4 percent and 10 percent. Estonia's and Latvia's rates are higher at 17.5 and 34.6 percent, correspond- ingly. For industrial products, Albania, the Czech Republic, the Slovak Repub- lic, Hungary, and Estonia have lower average rates of 6.6, 4.2, 4.2, 6.9, and 7.3 percent, respectively. 6. This is an extra Armenia-specific commitment to observe the provisions of the WTO Agreement. 7. Data show that a significant improvement in timeliness in VAT refunds began to take place in 2005. 8. Their design and implementation would require the attention of pol- icy makers. Some possible mechanisms are a rebate scheme on account, a deferred drawback, or a temporary admissions mechanism. 9. Freight in thousand metric tons stood at 33 percent of its 1997 level in 2003 and in million ton-kilometers at 26 percent. 10. According to this approach, a service can be provided only if it is explicitly permitted. 11. The fifth freedom refers the right of an airline from one country to land in a second country, pick up passengers, and fly to a third country where the passengers then deplane. This would allow, for example, a U.S. airline to stop en route to Armenia or fly beyond Armenia providing services at each stage. Chapter 6 Knowledge and Innovation T he creation, dissemination, and use of knowledge, together with the required strengthening of the competition and external inte- gration policies, must lie at the heart of the effort to ramp up technol- ogy in production. For the efficient use of knowledge, the four pillars that are widely accepted conditions are the following: an economic and institutional regime that provides incentives for the efficient creation, dissemination, and use of existing knowledge an educated and skilled population that can create and use knowledge a system of research centers, universities, think tanks, consult- ants, firms, and other organizations that can tap into the grow- ing stock of global knowledge, and assimilate and adapt it to local needs a dynamic information infrastructure that can facilitate the effective communication, dissemination, and processing of information Armenia is distinctive in having the following characteristics: A significant annual inflow into the labor force of energetic, moti- vated individuals who value knowledge and higher education. Armenia's private higher education (although of a predictably low quality) is one of the most vibrant sectors of the economy. A large stock of highly educated people, yet with largely obso- lete specialized skills. In the former USSR Armenia specialized in 117 118 Part I: Policies to Sustain Economic Growth R&D, and all the tribulations of the past 12 years notwithstand- ing, Armenia still boasts a critical mass of human capital that sustains a culture that values knowledge. However, this stock has been eroded recently owing to emigration, a low level of public spending on education, and delayed reforms in university education. A large and entrepreneurial diaspora which is as generous in philanthropic contributions to Armenia as (with few exceptions) it is reticent in business initiatives. Weak local entrepreneurship. Although there are some success- ful entrepreneurs (for example, in the software and jewelry sectors), they are not nearly numerous enough to change the economy. Clusters and value chains are not developing. There are of course the usual reasons of a weak investment climate and geographical isolation, but these do not appear to be the main problem. Armenian entrepreneurs and policy makers alike do not appreciate (and hence do not seek to improve) the value of intangibles (brand names, business reputations, marketing and managerial skills, networks, and so forth). This is where a large stock of educated engineers and scientists, with their attendant focus on assets that one can touch, turns to a disadvantage. Fragmentation of the policy debate. Traditions of collective action and public-private partnerships are also weak. The policy challenge lies in building on the strengths of the coun- try and addressing the critical weaknesses so that growth at higher productivity levels can be ensured. Here the major burden falls on the mobilization and recombination of the existing human capital, triggered by an initially modest investment in intangibles, such as mechanisms of knowledge and skill transfer from the diaspora to Armenia. The major contribution that policy analysis can make is to provide a fresh per- spective on problems, thereby inviting policy makers to change their mindsets and to identify and initiate knowledge-intensive initiatives. THE KNOWLEDGE ECONOMY IN THE INTERNATIONAL CONTEXT The following figures illustrate Armenia's knowledge readiness from the mid-1990s to the early years of this decade. Figure 6.1 demon- strates the relative performance of countries and regions in the Knowl- edge Economy Index (KEI), a composite index which measures the preparedness of the country for a knowledge-based development framework. The KEI is calculated by computing the average of the Knowledge and Innovation 119 Figure 6.1 Armenia and the World: Knowledge Economy Index 10 W. Europe 9 G-7 LEVEL Singapore United Rep. of Korea States Slovenia Oceania 8 Hungary Ireland G7 Poland Israel 7 ENTRY LEVEL Latvia Chile Estonia Slovak Rep. Czech Republic ecent)r 6 EAP ECA Russian Federation Belarus Costa Rica 5 COPING Armenia Argentina (most ECONOMIES LAC Romania 4 MNA Ukraine China Kazakhstan Mexico index 3 LAGGARDS Uzbekistan Turkey Georgia India 2 AFR SA 1 0 0 1 2 3 4 5 6 7 8 9 10 index (1995) Source: KAM, K4D, WBI, World Bank (www.worldbank.org/kam). performance scores of a country or region in all four pillars of the Knowledge Economy (KE). Overall, during the last five to seven years, Armenia improved rather marginally its performance in the KEI. In the knowledge map (Figure 6.1), as defined by countries' performance in the index, Arme- nia stands within the medium performers group, yet noticeably below the average of the ECA region. Armenia falls further behind OECD and EU members, while the performance gap between Armenia and its selected comparators is significant.1 In particular the gap with respect to Israel and Ireland is significant. These indicators should signal the sense of urgency for Armenia, which, despite its well-documented potential and impressive economic growth trends, does not seem to be maximizing the benefits from its competencies. Armenia is also falling behind other former socialist economies that have managed to com- pete better in a rapidly changing knowledge-based environment. Countries can be divided into four broad classes: Very low knowledge endowments--laggards. Relatively low endowments--coping economies. 120 Part I: Policies to Sustain Economic Growth "Accession club"--entry-level to KE--a class that is characteris- tic of upper middle­income economies starting to compete on knowledge and innovation, not on low labor costs alone. These are more advanced ECA and Asian countries. G-7 levels of knowledge endowments. While the criteria distinguishing one class from another are neces- sarily arbitrary, the messages for Armenia are quite clear: On the overall KE score, Armenia is among the coping econo- mies: it is within the same league as Russia, Ukraine, and Costa Rica and within a healthy distance from both laggards and entry- level economies. Yet Armenia's performance across the four KE pillars (economic incentive regime, education, innovation, and ICTs) is unusually unbalanced. On economic and institutional regime Armenia per- forms very well, on a par with recent EU entrants. In contrast, the ICT pillar is shockingly underdeveloped. Not only is Arme- nia squarely among laggards (the worst category, occupied by Sub-Saharan Africa, Albania, and the Central Asian republics of the FSU); its relative position has actually worsened significantly since 1995. This is all the more worrisome given its geographical isolation. Because of its landlocked status and unfriendly neigh- bors it should at least rank in the accession club on ICT. On education, Armenia still performs reasonably although lag- ging behind many Eastern European economies. However, the relatively good human capital is a heritage of Soviet times that has not translated into adequate innovation performance. The innovation pillar remains weak. Examining performance in the four KE pillars that define the aggre- gate KEI (Figure 6.2), Armenia performs poorly in the ICT pillar, an area in which it lost significant ground relative to the world. In abso- lute terms Armenia did improve its ICT indicators (explicitly shown in Figure 6.2) but the world on average--defined by the 121 countries in the Knowledge Assessment Methodology (KAM) sample--made a significantly larger improvement. Armenia's strongest pillar is its economic and institutional regime (EIR), which was the weakest in the mid-1990s; in this area the country demonstrated significant improve- ments and remains particularly competitive in the ECA region. In the education pillar, a traditionally strong area for the country, Armenia lost some ground, while in the innovation pillar the country regressed even more. Figure 6.3 illustrates how Armenia performed, over time, in each of the 12 variables that describe the four KE pillars, and therefore Knowledge and Innovation 121 Figure 6.2 Armenia: Performance in the Four Knowledge Economy Pillars, 1995­Most Recent Armenia (most recent and 1995), normalization group: all economic incentive regime 10 information innovation 0 infrastructure most recent 1995 education Source: KAM, K4D, WBI, World Bank (www.worldbank.org/kam). Figure 6.3 Armenia Knowledge Economy Index Scorecard Armenia (most recent and 1995) average annual GDP growth (%) Internet users per 10,000 10 Human Development Index computers per 1,000 tariff & nontariff barriers telephones per 1,000 regulatory quality (mainlines & mobiles) tertiary enrollment rule of law secondary enrollment researchers in R&D/mil. pop. adult literacy rate (% age 15 scientific & technical journal & above) patent articles/mil. pop. applications most recent granted by the 1995 USPTO/mil. pop. Source: KAM, K4D, WBI, World Bank (http://www.worldbank.org/kam). the aggregate KEI, plus two performance variables (GDP growth and Human Development Index [HDI]). In the EIR pillar, the country demonstrated the largest improve- ment in the ECA region, currently performing well above the regional average, behind Latvia and Hungary (Figure 6.4). Armenia signifi- 122 Part I: Policies to Sustain Economic Growth Figure 6.4 Economic Incentive Regime: Armenia and the World 10 Singapore 9 G-7 LEVEL Estonia Oceania W. Europe 8 Ireland Chile G7 United States 7 Hungary Costa Rica Israel ENTRY LEVEL Rep. of Korea Armenia Latvia Czech Rep. recent) 6 Poland Slovak Rep. Slovenia 5 COPING Mexico ECONOMIES (most EAP Turkey ECA 4 LAGGARDSGeorgia AFR MNA Romania LAC index 3 China Ukraine India 2 Russian SA Federation Argentina Kazakhstan 1 Belarus 0 Uzbekistan 0 1 2 3 4 5 6 7 8 9 10 index (1995) Source: KAM (http://www.worldbank.org/kam). Note: Armenia is an open economy with total trade representing currently 75 percent of GDP, which is still below the average of the ECA region. Armenia's exports of goods and services as a share of GDP--a solid indication of international competitiveness--is significantly lower than all comparator countries and the same income group average. In Armenia in 2002, despite the fact that exports of goods and services had significantly increased since 1995 and accounted for almost 30 percent of GDP, imports accounted for 47 percent of GDP, causing Armenia to carry over a significant trade deficit, much larger than any of its comparator countries. cantly improved its terms of trade by reducing tariff and nontariff bar- riers2 and also managed to improve significantly, relative to the world, its regulatory quality indicators by implementing policies friendly to markets and business. Rule of law indicators did not improve signifi- cantly and they are below the regional average levels, but still the country showed better improvement than the average of the KAM country sample, and therefore it demonstrates improvement relative to the world. Armenia is lagging behind all of its selected comparators in this pillar, with the exception of Russia, which is moving particu- larly slowly toward a modernized institutional phase. Figures 6.4 and 6.5 show several additional KAM indicators that describe performance in the pillar of economic incentive regime. The scorecard indicates that Armenia is characterized by low levels of gross Knowledge and Innovation 123 Figure 6.5 Armenia's Scorecard in the Economic Incentive Regime Armenia (most recent) gross capital formation (18.80) 10 5 tariff & nontariff 0 general government barriers (10.00) budget balance as % of GDP (-3.00) trade as % of GDP (75.00) Source: KAM (www.worldbank.org/kam). Note: Data for gross capital formation shown in the figure is the average of the period 1991­2001. domestic investment, the lowest among its comparators. On average for the years 1995­2002, Armenia spent around 19 percent of GDP in domestic investments--the least share along with Costa Rica--while the closest was Russia, 20 percent; and Latvia and Slovenia topped the list with 25 percent. Armenia on average for 1995­2002 spent a significantly smaller amount in domestic investments than, on aver- age, the groups of low- and middle-income countries. On governance and institutional quality variables, Armenia per- forms significantly worse than all other comparators and the average of the region, with the exception of Russia in rule of law and Israel in voice and accountability indicators. In particular, despite a national campaign against corruption, the high incidence of corruption con- tinues to affect business and the attraction of foreign investment. The Heritage Foundation (2004) states in its analysis that "bribery is wide- spread and is the most common form of corruption, especially in the areas of government procurement, all types of transfers and approvals, and such business-related services as company registration, licensing, and land or space allocation." The Bleyzer Foundation in a benchmark- ing analysis of 15 FSU countries on FDI driving conditions notes that Armenia is ranked thirteenth in the corruption level index, whereas in the overall composite index it is ranked sixth, behind Estonia, Latvia, Lithuania, Kazakhstan, and Russia (Bleyzer 2002).3 124 Part I: Policies to Sustain Economic Growth Furthermore, the size of the shadow economy in Armenia, which is equal to 45 percent of GDP, is significantly larger than that in some of its comparator economies and the average for the region (Table 6.1). This reveals the poor institutional capacity and high incidence of corruption in Armenia, combined with the increasing tax burden and social security payments. The size of the shadow economy in Armenia is a prohibitive factor in fiscal revenue generation, which could cre- ate a vicious cycle with tax base erosion (resulting in higher taxes), the worsening of fiscal constraints (Economist 2003), and ambiguous effects on private sector development and the quality of products and services. Unless urgent and radical reforms transform the effectiveness of the governance and institutional capacity of the country, Armenia will be facing competitiveness challenges that will be hard to meet. In education, a pillar in which Armenia has a strong tradition, the country lost some ground relative to the world and remains weak rel- ative to the regional ECA average (Figure 6.6). Enrollment in second- ary education has fallen significantly in the post-transition period and remains well below the (ECA) regional average. However, enrollment in tertiary education improved significantly but still remains at very low levels by regional standards.4 The vast majority of ECA econo- mies, and all selected comparators (with the exception of Costa Rica), outperform Armenia in the KAM variables used to define aggregate performance in the education pillar. Figure 6.7 isolates Armenia and presents the available set of vari- ables used in the KAM to define performance in the education pillar. It is striking to realize, relative to the large availability and potential of educated human capital in the country, how little Armenia spends on education--less than 3 percent of GDP (2000 data). The significant Table 6.1 The Size of the Shadow Economy (as percent of GDP) Country 1990­93 2000­01 Armenia 40.1 45.3 Estonia 34.3 39.1 Ireland 15.4* 15.7** Russian Federation 27.8 45.1 Slovenia 22.9 26.7 Average Central and Eastern Europe 23.4 29.2 Average FSU 32.9 44.8 Average of 21 OECD countries 15.7* 16.7** Source: Schneider 2002. *1994­5; ** 2001­2. Knowledge and Innovation 125 Figure 6.6 Education: Armenia and the World 10 W. Europe United States Poland 9 G-7 LEVEL Slovenia Ukraine Oceania Latvia 8 Russian Federation G7 Czech Rep. Ireland Argentina 7 Rep. of Korea ENTRY LEVEL Slovak Rep. ECA Israel Estonia Romania Belarus recent) 6 Chile Hungary EAP Kazakhstan 5 Mexico Georgia (most COPING LAC Uzbekistan 4 ECONOMIES Singapore MNA Costa Rica Armenia Turkey index 3 LAGGARDS China 2 India SA AFR 1 0 0 1 2 3 4 5 6 7 8 9 10 index (1995) Source: KAM (http://www.worldbank.org/kam). Figure 6.7 Armenia's Scorecard on Education Armenia (most recent) adult literacy (% age 15 and above) (38.50) 10 public spending on 5 secondary enrollment education as % of (73.28) GDP (2.90) 0 life expectancy at tertiary enrollment birth, years (74.10) (20.16) primary pupil teacher ratio, pupils per teacher (18.50) Source: KAM (www.worldbank.org/kam). 126 Part I: Policies to Sustain Economic Growth gap in education spending between the ECA regional average (4.6 percent of GDP) and Armenia is particularly alarming. Based on the KAM variables that define the innovation pillar, which in the mid-1990s was the strongest pillar of the country, Armenia lost significant ground relative to the world, and its performance currently falls below the ECA regional average (Figure 6.8). It is evident that Armenia lost a significant part of its stock of researchers (brain-drain) in the diaspora, while the number of scientific and technical publica- tions has declined over the years. Patent activity is minimal and has remained rather stagnant throughout the years. Although close behind Costa Rica and Chile, Armenia lags behind all selected comparators. In Figure 6.9, basic KAM indicators using the most recent available data are presented describing the performance of Armenia in the inno- vation pillar. One of the weakest indicators for Armenia is spending on R&D (0.2 percent of GDP) whereas the average for the lower-middle- income group is close to 0.9 percent (Figure 6.9). For the ICT pillar, in the variables that describe the availability and penetration ratios of these technologies, Armenia scores dramati- cally below the world and the ECA regional average, indicating the country's weakness in keeping up with regional and global technol- ogy penetration and usage trends (Figure 6.10). ICT is the country's weakest pillar. In absolute terms, however, some improvements in Figure 6.8 Innovation: Armenia and the World 10 United States 9 G-7 LEVEL Israel Singapore G7 Rep. of Korea W. Europe 8 Oceania Slovenia Ireland Hungary Russian Federation 7 ENTRY LEVEL Czech Rep. Estonia Slovak Rep. Argentina Belarus recent) 6 Latvia ECA Ukraine Romania Poland 5 COPING Mexico Armenia ECONOMIES Georgia Turkey EAP Chile (most China Costa Rica 4 Kazakhstan MNA India Uzbekistan index 3 LAGGARDS LAC AFR 2 SA 1 0 0 1 2 3 4 5 6 7 8 9 10 index (1995) Source: KAM (http://www.worldbank.org/kam). Knowledge and Innovation 127 Figure 6.9 Armenia's Scorecard in Innovation Armenia (most recent) gross foreign direct investment as % of GDP (3.70) high-tech exports as % of 10 science & engineering enrollment ratio manufactured exports (4.00) (% of tertiary level students) (29.00) patent applications granted by the 5 researchers in R&D (4,971.00) USPTO/mil. pop. (0.32) 0 patent applications granted researchers in R&D/ by the USPTO (1.00) mil. pop. (1,603.55) scientific and technical journal total expenditure for R&D articles/mil. pop. (37.36) as % of GNP (0.18) scientific and technical manufacturing trade as % journal articles (142.00) of GDP (30.68) Source: KAM (www.worldbank.org/kam). Figure 6.10 ICT: Armenia and the World 10 Rep. of Korea Singapore 9 G-7 LEVEL Slovenia G7 Israel 8 Slovak Rep. Estonia Ireland United 7 Chile ENTRY LEVEL Latvia Czech Rep. States Hungary Oceania Argentina Costa Rica W. Europe recent) 6 EAP Poland COPING Belarus Romania Turkey 5 ECONOMIES MNA Russian Federation (most Mexico China LAC 4 ECA Georgia Ukraine index 3 LAGGARDS Armenia Uzbekistan Kazakhstan 2 AFR India SA 1 0 0 1 2 3 4 5 6 7 8 9 10 index (1995) Source: KAM (www.worldbank.org/kam). Internet usage and computer penetration ratios were achieved, but the volume of these improvements was much less significant than those occurring globally. In Armenia the current levels of Internet users per 10,000 people and computers available per 1,000 people are among the lowest in the region and globally. The availability of telephones and 128 Part I: Policies to Sustain Economic Growth Figure 6.11 Armenia's Scorecard in ICT Variables Armenia (most recent) telephones per 1,000 (mainlines & mobiles) (151.50) 10 Internet users per 10,000 (184.12) main telephone lines per 1,000 (139.80) 5 Internet hosts per 10,000 (7.50) mobile phones 1,000 (11.70) 0 daily newspapers per 1,000 (5.00) computers per 1,000 (9.20) radios per 1,000 (225.00) TV sets per 1,000 (230.00) Source: KAM (http://www.worldbank.org/kam). mobile phones per 1,000 people is also very limited--an additional element indicating the profound weaknesses and state of emergency in the telecommunications infrastructure of the country. The gap in performance between Armenia and its selected comparators is tre- mendous and is apparently widening. Armenia's lack of growth is paradoxical. There is a high potential of knowledge utilization that shows tantalizing promise (for example, in software and certain enterprises of heavy industry) yet remains largely untapped. Figure 6.11 shows Armenia's performance relative to the world in ICT variables. POLICIES TO STRENGTHEN KNOWLEDGE-BASED COMPETITIVENESS A pragmatic agenda for change will entail focusing on bottom-up entry points (immediate policy agenda), then scaling them up to assure coordination and concerted action (medium-term policy agenda), and then moving to major reforms (longer-run policy agenda). Finland, Ireland, and the Republic of Korea are the best-known, best-practice exemplars of concerted action; they have engineered successful transitions to knowledge-based economies. In all of these cases, national economic crises compelled diverse actors to define and implement a new agenda through explicit or implicit national agreements on goals and mechanisms to move forward. The crises also prompted policy makers and private sector leaders to lengthen Knowledge and Innovation 129 the time horizon of the policies adopted. Thus, Nokia--Finland's first mover toward an innovation-based economy--dramatically increased R&D investments. Finland responded by increasing public R&D and transforming the innovation system to fit business needs. Members of Parliament took courses and went on study tours demonstrating the need for the new agenda. National public innovation organizations played a crucial role by transforming technology into businesses and assuring a critical mass of demonstration cases. Ireland also exemplifies a successful combination of top-down and bottom-up policies. It made an investment in education and R&D infrastructure in the 1980s, followed by drastic policy changes begin- ning in 1987. To complement its top-down policies, Ireland instituted pragmatic bottom-up programs, including regional partnerships to mitigate high unemployment and a program to expand national- supplier linkages from FDI. Korea's powerful and shared national vision--from which impetus a private sector champion emerged-- was followed by effective government coordination (see Box 6.1). Three lessons from the discussion above are relevant for Armenia. First, simple institutional recipes do not exist for concerted action. Armenia will need to devise its own recipe for a KE. Given its great regional diversity, Armenia's regional and state-level policy initiatives would be a key entry point for a knowledge-based economy. Armenia has already advanced somewhat in that direction. Second, the experiences of Korea and Finland indicate that even when major crises call for urgent economic transitions, countries must "prepare the bases." This essential preparatory stage is seen in the ini- tial Vision Korea Report. It is seen in Finland's major effort to facilitate and accelerate business R&D. Third, although major reform efforts from the top are vital, they may not provide the all-important impetus for transformation. Con- certed effort must evolve. Bottom-up experiments in Armenia, some of which are already under way, must mature. These transitional stages then proceed to concerted efforts (the Korean knowledge strategy is one example). Drawing on a diversity of best practices, we suggest that Armenia construct and implement a strategy to move toward knowledge-based competitiveness in three stages: Immediate agenda (2007­08): Begin massive awareness-building and initiate pilot/demonstration projects. Medium-term agenda (2008­10): Create a springboard for major reforms by assuring major improvements in the investment cli- mate, strengthening stakeholders for reforms, and proceeding with a private sector-led shared vision "Armenia 2025." 130 Part I: Policies to Sustain Economic Growth Box 6.1 Korea's Transition to a KE: Bottom-up Initiative Leads to Government Action In 1998, Korea officially launched a national strategy to move to a knowl- edge-based economy in the wake of a financial crisis. The initial impe- tus came from the private sector--the Maeil Business Newspaper--which concluded in 1996, even before the crisis, that there was an urgent need for a more coherent vision of the future of the Korean economy. This newspaper launched the "Vision Korea Project" as a national campaign in February of 1997, and developed the first Vision Korea Report. Eventually, the government--the Ministry of Finance and Econ- omy--became the main champion of the KE policy agenda. The Korean Development Institute was a so-called system integrator and coordi- nated the work of a dozen think tanks. A joint World Bank and OECD report provided a framework, outlining concrete steps for reforms in the various policy domains. Progress was monitored closely. This was a crucial step in identifying and addressing any inertia or resistance, as for example, with educa- tion. Korea's knowledge strategy of April 2000 evolved into a three- year action plan for five main areas: information infrastructure, human resources, knowledge-based industry, science and technology, and the elimination of the digital divide. To implement the action plan, Korea established five working groups involving 19 ministries and 17 research institutes, with the Ministry of Finance and Economy coordinating the implementation. Every quarter, each ministry submits a self-monitoring report to the Ministry of Finance and Economy, which puts out an inte- grated report detailing progress. The midterm results and adjustments to the plan are sent to the executive director of the National Economic Advisory Council, which reports on the progress of the implementa- tion and gives an appraisal of the three-year action plan to its advisory members. Source: Author's own elaboration. Long-term agenda (2008 onward): Enact major reforms that would transform and create world-class innovation capacity, education systems, and ICT infrastructure (Table 6.2). Table 6.2 outlines a medium-term agenda (establishing a spring- board for the KE) and a longer-term agenda (creating a world-class KE infrastructure). We view the decade from 1995 to 2003 as a stage of building foundations. Armenia's accession to the WTO in 2003 was an important event signaling that Armenia had a certain macroeconomic and socioeconomic stability and had created the basic institutional Table 6.2 Sequencing of the Armenia KE Policy Agenda Stages of economic reform Drivers Examples of and growth Major constraints of growth Thrust of government policy policy initiatives Benchmarks Building Sustainability of Remittances Assuring macro and social Infrastructure Signaling event: foundations macroeconomic and other stability projects Accession to WTO (2003) 1995­2003 stability and transfers Some initiatives to improve Creation of business Other indicators: Investment share market reforms from abroad innovation climate council in GDP, private investment in GDP Infrastructure (reasonably high) and services Level of knowledge-based exports (very low) Business share in innovation (practically nil) Establishing a Dearth of role Increasing Assuring a critical mass of Competitive grant Signaling events: springboard models and share of stakeholders for reforms schemes to enhance A multinational establishes for KE, stakeholders for merchandise through a two pronged business innovation knowledge-intensive operations in 2005­8 reforms (self- exports; strategy: Armenia Knowledge Creation of made start-up growing role Top-down: dramatic innovation council Elimination of ArmenTel monopoly and spin-off of services reduction of administrative entrepreneurs) Some skilled emigrants come barriers to growth back and become successful and Bottom-up: facilitation entrepreneurs of private sector-driven Rising share of business R&D and Innovation "centers of excellence" merchandise exports Major Human capital Skill-intensive Major overhaul Wide-scale Significant return migration of reforms: constraint: exports of education, ICT introduction of highly skilled creating inadequate stock infrastructure, and income-contingent Robust Robust knowledge-intensive world- and flow of innovation systems scheme to finance internal clusters are established 131 class KE technical skills private higher demand infrastructure, education Inadequate ICT 2008­15 infrastructure Source: Elaboration by the author. 132 Part I: Policies to Sustain Economic Growth foundations for a market economy. We argue that the time has come for a new stage of policy: assuring a critical mass of stakeholders for reforms. That, in turn, would imply a two-pronged strategy: a top-down approach: a dramatic reduction of the administrative barriers to growth and a dramatic improvement in the invest- ment climate a bottom-up approach: the facilitation of private sector-driven "centers of excellence" in innovation, enterprise upgrading, edu- cation, and ICT. This policy stage should be characterized by more strategic and proactive government policies to improve innovation, the education systems, and the ICT infrastructure. By proactive, we do not mean a sectoral industrial policy of "picking winners." Rather, the focus should be on functional interventions that are open to all eligible stakeholders and that are designed to accelerate existing policy trends, rather than creating policies from scratch. Examples of policy initia- tives in this vein include competitive grant schemes to enhance busi- ness innovation (more on this is given below) and the creation of an innovation council to assure linkages among higher education, R&D organizations, and the enterprise sector. Benchmarks and signaling events of the completion of this stage of reform could be as diverse as the following: A multinational establishes knowledge-intensive operations in Armenia. The ArmenTel monopoly is eliminated. Some skilled emigrants come back to Armenia and become suc- cessful entrepreneurs. This is a rising share of business R&D in the overall R&D budget and of merchandise exports in overall exports. The achievement of these benchmarks would signal the formation of a critical mass of stakeholders that would allow Armenia to engage in major and quite painful reforms in innovation, in the enterprise upgrading system, and in ICT infrastructure. We will not discuss in detail the long-run agenda for change. What follows focuses on the "what" and "how" of the medium-term agenda (establishing a spring- board for a KE). To summarize, the sequencing of the transition to a knowledge- based economy in Armenia can be conceptualized as focusing on bottom-up entry points (the immediate policy agenda), then scaling them up to ensure coordination and concerted action (the medium- Knowledge and Innovation 133 Figure 6.12 Virtuous Circle of Growth and Reforms top down momentum Intermediate agenda Medium-term agenda Longer-term agenda demonstration projects critical mass structural reform sense of urgency of changes bottom up momentum Source: World Bank staff. term policy agenda), and then moving to major reforms (the longer-run policy agenda). The art and craft of policy making is about sequencing the various horizons of a policy agenda in a virtuous circle of growth and reforms. A pragmatic agenda to get around the many institutional rigidities that Armenia faces includes the following steps: (i) create the momentum for change by fostering stakeholder awareness; (ii) attain consensus on tackling some of the key obstacles at the national level (to enhance the demand for an institutional change); and (iii) to move ahead with concrete, manageable, bottom-up approaches that can serve as demonstration projects to move the larger agenda (Figure 6.12). MEDIUM-TERM AGENDA: ALLEVIATION OF CRITICAL CONSTRAINTS The arguments about the sequencing of reforms discussed in the pre- vious section have been confirmed by experienced policy observers. For example, Dani Rodrik's paper (2004) argues that the key to growth is not getting all or most institutions right at once, but rather overcom- ing the chief bottleneck to raising growth by, say, 2 percentage points a year, and using the proceeds of this improvement to overcome the next bottleneck--and so on. There is a kind of `bootstrapping reform' strategy which provides useful insight into the "how to" of reforms. 134 Part I: Policies to Sustain Economic Growth As is evident from the analysis in Chapter 1, two major immediate constraints from the perspective of knowledge-based competitiveness are the following: an extremely weak and fragmented innovation system, includ- ing lack of linkages among the productive sector, the universi- ties, and the research institutes the low quality and high prices of ITC infrastructure From a longer-term perspective, human capital (particularly techni- cal and managerial skills) is a major constraint. Table 6.3 summarizes the critical constraints from the perspective of the KE, and also the relevant medium-term policy agenda (to be discussed below) for alleviating the constraints. INNOVATION SYSTEM: MAKING INNOVATION RELEVANT FOR BUSINESS An innovation system consists of a network of organizations, rules, and procedures that affects how a country acquires, creates, dissemi- nates, and uses knowledge. Key organizations for the creation of knowledge include universities, public and private research centers, and policy think tanks. Private firms are at the center of the innova- tion system. If the private sector has little demand for knowledge, the innovation system cannot be effective. Effective linkages between R&D and industry linkages are vital to transform knowledge into wealth. Therefore, networking and interactions among the different organiza- tions, firms, and individuals are critically important. The intensity of these networks, as well as the incentives for acquiring, creating, and sharing knowledge, are influenced by the economic incentive regime in general. The innovation system framework, as it has been applied to a vari- ety of studies of OECD countries, relies too much on innovation in the sense of the development of radical or incrementally new knowledge. Traditional measurements for an innovation system include indicators of expenditure on R&D, activity in high-technology sectors (biotech- nology, ICT), patenting activity (number, intensity), and researchers per 10,000 population. These indicators proxy the ability to generate new knowledge. However, they are not particularly helpful in under- standing how a traditional, low-tech manufacturing firm can learn to upgrade its capabilities to compete in a more knowledge-based econ- omy. Rather, these indicators are very often just the top of a pyramid (see Figure 6.13), the bulk of which comprises firms, mostly SMEs, for Table 6.3 Specific Policy Initiatives as Entry Points to Address Systemic Constraints Relevant best practices System Main objectives Specific initiatives (discussed in Chapter 3) Innovation system Enhance linkages among Competitive grant scheme to U.K. Teaching Company Scheme productive sector, universities, promote business R&D and and science organizations training Seed support to venture capital fund Education system Reform curriculum of basic, Technology/design facility on Competence center models secondary, and higher education the basis of major university (in design stage in Estonia) Distant education project Knowledge collaboration with diaspora Establish National Innovation Council to bring together and education, R&D organizations, and industry Innovation ICT infrastructure Improve quality of ICT services Institutional strengthening of Liberalization policies in ICT and reduce their costs public utility regulator infrastructure in some Central European countries Source: Elaboration by the author. 135 136 Part I: Policies to Sustain Economic Growth Figure 6.13 The Pyramid of Learning Capabilities of Firms Research and technology development: advanced firms Very rarely present, mostly large firms R&D Design and engineering: technologically competent firms technology Capabilities rarely present in SMEs upgrading reverse engineering Technician and craft skills and capabilities: minimal technology firms In SMEs, strong skills sometimes present, technology acquisition though key skills often absent or weak assimilation Basic operating skills and capabilities: survival-oriented enterprises technology use In SMEs, often weak, with and operation limited and irregular upgrading Source: Adapted from Intarakumnerd et al. 2002. which the major issue is the acquisition of basic skills in marketing, design, engineering, and other operational areas rather than technol- ogy upgrading and R&D. Because of this, the traditional innovation system approach might be applicable to a very limited subset of the economy, such as firms in the export-oriented sectors. It might also be useful for setting long- term goals and objectives, in terms of creating the elements missing in the traditional approach. Table 6.4 flags a range of policy interven- tions--which should be considered as a menu of options, not all of them immediately relevant for Armenia. Armenian policy has a narrow view of technological development, which is equated with R&D. The underlying model of technology development describes firms only as the demand side, relative to the supply side that is provided in R&D institutes and universities. Policy considers that the key to reducing the technology gap is not in the technology activities of enterprises themselves but through the expansion of intermediary institutions such as technology centers Table 6.4 Matching Policies to Capabilities: A Range of Instruments to Support Innovation Type of enterprise Policy objectives Instruments and interventions Survival- To build basic competitive business advisory and support services--SME and micro-enterprise oriented capabilities by fostering support agencies enterprises awareness of scope and benefits facilitation of access to finance (including micro-finance) of innovation management and skills development Minimum To foster competitiveness by support for business development, diversifying customer base technology introducing basic innovation product diversification and quality improvement firms skills and encouraging adoption management and skills development and application of new ideas Internet-based information services technology awareness and marketing support for technology adoption and adaptation projects cluster-based approaches to stimulating innovation Technologically To support market development business development, exports market support competent and entry into global value Internet-based information services enterprises chains by fostering strategic technology transfer support Knowledge alliances and certain in-house incubators and technology parks innovation capabilities linkages with academic researchers laboratory services and metrology services consultancy and technical assistance support, such as on and commercialization, intellectual property rights, licensing, patenting supplier development and linkage promotion programs Innovation Advanced To facilitate moving up global support for participation in international R&D networks (such as the enterprises value chains by upgrading in- EU 6th Framework Program) house innovation capabilities technology and other innovation-based spin-offs and strategic alliances university-industry collaboration support for commercialization Diffuse experience of innovation 137 development of vibrant venture capital industry leaders as role models for the encourage participation of national innovation leaders in national rest of the economy advisory bodies, technology foresight, and cluster processes Source: Elaboration by the author. 138 Part I: Policies to Sustain Economic Growth and science and technology (S&T) parks. The problem of technology development is reduced to the issue of the commercialization of R&D results. Treating industry only as demand side is quite misleading because industrial firms not only generate the demand for industrial technol- ogy but they account for a very large part of the supply side as well. In fact, most of the technology-generating capabilities are located in the industry (that is, in firms themselves, not in extramural organiza- tions such as S&T parks and R&D institutes). Business enterprises fund between 50 and 60 percent of the general expenditures on research and development (GERD) in North America, the EU, and the Nordic countries, and they perform between 60 and 70 percent of the GERD. Thus, the issue for Armenia is how to increase R&D in the business sector, not outside of it. In addition, using only R&D can be misleading as a large contribution to technology development is made by types of technical change that do not involve formally organized R&D at all. Innovation surveys from the EU and Central European economies show that the key source of important information for innovation comes from enterprises themselves or from partners in value chains (suppliers, buyers). Infrastructure institutions are actually marginal as a direct source of information for innovation. The importance of non- R&D technology development suggests that it is inappropriate to nar- row down innovation policy to just R&D policy and to bridging insti- tutions. In fact, infrastructure institutions are much more important as sources of knowledge and skills carried by people who move among universities, R&D institutes, and firms, or between firms. Information services, training, and standard services as well as problem solving and R&D in all forms of people-centered types of output are the key function of extramural technology institutes. One of the most impor- tant roles played by R&D in public technology institutes is to contrib- ute flows of people into the technological activities of industry. In Armenia, as in many Eastern European and CIS countries, there seems to be a dearth of measures that stimulate firms to undertake their own technology development. Policy is focused much more on meas- ures that support institutions in undertaking S&T activities on behalf of industrial firms. Within the innovation policy, there are no policy measures that support and facilitate actions by firms themselves. In a nutshell, Armenia should avoid an exclusive focus on support- ing technology institutions in a supply-driven approach, with support to extramural institutions rather than to firms. The balance should be corrected between a focus, on the one hand, on directly strengthening the technology development capabilities and activities of firms, and on the other hand, building and strengthening various kinds of tech- Knowledge and Innovation 139 nology development and transfer institutions. Currently, there is not a system that would support firms to advance from technology use and maintenance to technology development and creation. Mechanisms to support technology transfer institutions (such as S&T parks) are an important component of the policy system. However, a major emphasis in policy development should be on mechanisms that assist and empower firms to make their own investments into technology absorption and development. Within that firm-centered approach to policy, increasing emphasis has to be given to stimulating and facilitating various forms of collective activity involving groups of firms. These groups may be established industry associations, or less formally structured groups organized around value chains, or clusters of firms in related industries. Among instruments for financing innovation in Armenia, grant- based mechanisms are notably absent. These mechanisms commonly provide grants to firms which undertake particular kinds of techno- logical activity (such as R&D, design, technological or managerial training, engagement of consultants, employment of qualified scien- tists and engineers, and many others). In most cases, the grants cover a defined proportion of the costs of the specific activity. ICT INFRASTRUCTURE: ASSURING ENTRY OF NEW SERVICE PROVIDERS The beginning of this chapter documents a very low level of develop- ment for ICT infrastructure. Figure 6.1 shows that Armenia performs just below the average for economies of the FSU, which is hardly satisfactory given Armenia's peculiar geographical position and aspi- rations to develop a competitive ICT cluster.5 This is mainly attribut- able to one single problem: the ArmenTel monopoly, which has been addressed in earlier chapters. FORMATION OF HUMAN CAPITAL: ENHANCING EDUCATION­INDUSTRY LINKAGES Many problems in the education sector, especially in the school sys- tem, are well understood and are being addressed through existing and planned reforms. This section will focus in more detail on the education­industry linkages. Chronic underspending. Armenian public spending on education is 2.8 percent of GDP, well below the OECD average of 5 percent and the 140 Part I: Policies to Sustain Economic Growth rate in other transition economies. Private spending is also low (less than 0.5 percent) as is education as a proportion of public spending (11 percent). Inefficiency. A dramatic fall in the school-age population since inde- pendence, plus only modest reductions in staffing levels, has resulted in staffing ratios which are low in international terms and, more important, unsustainable in the Armenian context. For example, there are only about 11 pupils for every full-time equivalent teacher. The Medium-Term Expenditure Framework (MTEF) proposes increasing these overall ratios to one teacher for every 16 students. Lack of relevance. There has been little reform of the curriculum, the assessment methods, or the teacher training since independence. The main challenge is the transition from a teacher-centered to a stu- dent-centered learning approach. Reforms supported under the EQ&R project will be a major step toward meeting this challenge. Equity. Enrollment in basic education continues to fall, and is now below 85 percent. Low levels of public spending have resulted in increasing levels of informal payments in "free" basic education. In higher education, expansion has been mainly in private institutions and fee-paying places at public institutions. While this has increased access--though still for only 16 percent of the cohort--public spend- ing overwhelmingly favors the rich, since state scholarships are given mainly to those who score highest on the entrance examination. Governance and management. One of the biggest and most difficult challenges Armenia faces is to move away from the top-down system of management, in which the Ministry of Education passes directives that are then uniformly applied to all institutions and in which the flow of information is one way and is used to control rather than empower local actors. Tertiary education. A major stakeholder conference was held in November 2002, organized during the preparation of SAC V. The equity concerns have been mentioned above. The lack of quality assurance mechanisms (accreditation, inspec- tion, and so forth) means that students and other stakeholders cannot choose better institutions and courses, and good-quality private insti- tutions are undermined by a small number of fly-by-night operators. Investment is low or nonexistent, resulting in poor quality, outdated teaching environments, and an almost complete absence of interna- tional level research. Knowledge and Innovation 141 EDUCATION­INDUSTRY LINKAGES Historically, Armenia has had a highly educated population, and educational attainment figures have remained remarkably high since independence despite the low levels of spending. However, the skills and knowledge that individuals have acquired have become increas- ingly obsolete in the labor market. Much of the education would have been fact-based information learned by rote; and, even if the educa- tion were more relevant to the labor market, high levels of unem- ployment, underemployment, and informal employment would have eroded skills and knowledge. Older adults, though they are more likely to be employed, would also have worked for longer periods in state-owned, static industries. The story of a private company Lycos illustrates the main issues and options for enhancing education industry linkages. Lycos is a fully owned subsidiary of the German-based Lycos Europe. Lycos Armenia is integrated into Lycos Europe's network of competence centers in Gutersloh, Copenhagen, Hamburg, Munich, Paris, and Stockholm for the development of its core services: search, communications, commu- nities, and shopping. Lycos Armenia is in the software development business (Web posting, e-mail, chat room communities, e-commerce) and provides technical support to the Lycos Europe portal. Impor- tantly, the company works closely with two universities in Yerevan and provides grant funding of about US$500,000 for a two-year bach- elor's degree program in Internet computing designed primarily by Lycos staff. Sixty to 70 percent of the lectures are also given by Lycos staff. Throughout the program internships are also offered in the Lycos offices in Armenia and elsewhere in Europe. The top graduates of the program are usually hired by the company. The fact that a private company establishes its own training pro- grams and, more important, does this in collaboration with major Armenian universities, is promising. This indication of the effective demand of private sector and private sector initiative is a foundation on which a coherent private-sector-led system of lifelong learning can be built. The following options for expanding and accelerating the existing education-industry linkages could be considered: Distance learning as a pilot project (Box 6.2). The advantage of this potential distance learning project is that it would bring together the ICT and education dimensions. As Box 6.2 out- lines, the Armenian diaspora can become involved in a new and productive way. In addition, as worldwide experience of distance learning indicates, these projects are easily scaled up and expanded when successful. 142 Part I: Policies to Sustain Economic Growth Box 6.2 Distance Learning as a Potential Pilot Project to Enhance Education­Industry Linkages Distance learning could be a low-cost opportunity for Armenia to accel- erate the transfer of global knowledge and drastically upgrade the qual- ity of teaching in its universities. For a landlocked, remotely located country, the modern technology could provide the following group of primary benefits: Access to high-caliber professors and lecturers, who would initially demonstrate how the core modern curricula should be delivered to students and therefore would greatly contribute to the training and retraining of trainers (local professors). The availability of various professional talent in the diaspora and the existence of an established professional diaspora network would simplify the future mobilization of potential participants and could further reduce project costs (many diaspora members may be ready to supply such lecturing on a pro bono basis). Recent examples from Turkey and Thailand confirm the feasibility of such an approach. Online access to modern experimental facilities and academic libraries. Economies of scale, including low-cost dissemination/sharing of popular courses among various local universities and training centers. As with many other collective diaspora initiatives, the distance learn- ing project, especially in the area of engineering, is likely to lead rather quickly to the second generation of (indirect) benefits. As experience from other countries suggests, professionals participating in advanced educational projects abroad tend to be eager to launch new business ventures with their local partners and frequently with their former stu- dents. On the parallel track, collective efforts of diaspora activists in the area of university education has the potential to evolve gradually toward more business-oriented projects, undertaken basically by the same group of initial diaspora sponsors, such as those associated with university business incubators. Source: Elaboration by the author. Establishment of a Skills Development Fund to encourage enter- prise sector training, subject, of course, to a thorough cost-benefit and fiscal priority assessment. Knowledge and Innovation 143 SUMMARY AND KEY RECOMMENDATIONS Armenia enjoys various advantages that should lead to future growth that is based on innovation and the skilled use of knowledge-inten- sive assets: it has a core of well-educated and motivated labor, the incentive regime is in place, and there are international linkages. But local entrepreneurship has been weak, and ICT is handicapped by high telecommunications costs and low R&D spending. In addition to the reform in competition discussed earlier, the government could facilitate private sector-driven centers of excellence in innovation and enterprise upgrading. The government could also pursue reforms in education, infrastructure, telecommunications, and aviation to take explicit account of the need to support ICT and innovation. business advisory and support services--SME and micro- enterprise support agencies facilitation of access to finance (including micro-finance) management and skills development Internet-based information services support for technology adoption and adaptation projects cluster-based approaches to stimulating innovation support for participation in international R&D networks (for example, the EU 6th Framework Program) encouragement of participation of national innovation leaders in national advisory bodies, technology foresight, and cluster processes NOTES 1. In figures like the one presented in Figure 6.1 the reader should note the following points. On the horizontal axis the 1995 KEI scores are plotted. On the vertical axis the KEI scores for the most recent year are plotted. The more advanced KE performers plot in the northeast quadrant of the graph, while the weaker ones plot in the southwest quadrant of the graph. Both the position along the 45-degree line and whether a country plots above or below the line are significant. The countries or regions that are plotted below the line indicate a regression in their performance relative to where they were in 1995. The countries or regions that are marked above the line signify an improve- ment in their position in the latest period compared to their position in 1995. Those countries that are plotted on the line have maintained their relative position over the two periods. Each country's performance as depicted in the figure is relative to the performance of the total country sample included in the KAM (121 countries). 144 Part I: Policies to Sustain Economic Growth 2. This reflects 2003 data from the Heritage Foundation. Nevertheless, the Heritage Foundation states the following in 2004 Index of Economic Freedom: "In 2001, according to the World Bank, Armenia's weighted average tariff rate was 2.5 percent, up from the 1.9 percent reported in the 2003 Index of Economic Freedom by the Heritage Foundation." The U.S. State Department reports that most imports are free of prohibitions, quotas, or licensing, but businesses complain about "cumbersome procedures [and] bribes solicited by customs officials." Based on new evidence of customs corruption, Armenia's trade policy score is 1 point worse in this year (2004). 3. In this analysis 15 FSU countries are ranked by the following FDI driv- ing elements: liberalization and deregulation of business activities, stability and predictability of the legal environment, corporate and public governance, liberalization of foreign trade and international capital movements, financial sector development, corruption level, political risk, country promotion and image, and targeted investment initiatives. 4. Based on statistical information management and analysis (SIMA) data for the year 2000, Armenia is performing better in tertiary enrollment rates than Albania, Tajikistan, and Turkey. 5. This section relies on the Public-Private Infrastructure Advisory Facility (PPIAF) presentation, "Sector Overview and Review of International Experi- ence, Identification of Bottlenecks and Recommended Roadmap to Develop a Rural Telecommunications Strategy in Armenia" (PPIAF 2004). Part II: Detailed Analysis Chapter 7 Growth Analysis from the Perspective of Employment Generation and Poverty Reduction SUBSECTORAL GROWTH, INVESTMENTS, AND EXPORTS Sectoral Growth This chapter1 discusses the economic transformation that is associ- ated with job creation outside of subsistence agriculture and financial deepening--processes that took place in new EU member countries. The transformation has been slow in Armenia, but has started to pick up recently. Agriculture's contribution to GDP in Armenia continues to be high at 25­30 percent and its contribution to employment is more than 45 percent, well above the levels of 4.5 percent and 5 per- cent in new EU member states. Agriculture is concentrated in low value-added products (dairy, meats, and grains). The services sector, which is concentrated in public administration and trade, generates 35­37 percent of GDP in Armenia, compared to more than 60 percent in new EU member countries (Figure 7.1). Growth was led by construction and industry (including manufac- turing), mainly through power production, food processing, and bev- erages, and in part by services, through trade and catering (but with services growth being notably slower than in other fast-growing transi- tion economies) and mirrored export and investment performance. Sectoral contributions to GDP have changed dramatically over the past decade. The most striking feature has been the growth of the services sector and construction at the expense of industry (Table 7.1). Changes in relative prices had a dramatic impact on agriculture (that is, a decline in agriculture terms of trade since the 1998 Rus- sian crisis). Real agricultural prices have been declining until recently, while real prices in services (led by the utility price adjustment) and 147 148 Part II: Detailed Analysis Figure 7.1 Armenia and the Slovak Republic, Employment by Sectors, 1994­2004 100 80 60 cent per 40 20 0 1994 2003 1994 2003 Armenia Slovak Republic employment in agriculture employment in industry employment in services Source: WDI. Table 7.1 Sources of GDP Growth in Armenia, by Sector Contribution to GDP Sector Share in GDP (percent) US$ million Agriculture 24.8 531.3 Construction 13.9 298.6 Diamonds 1.9 40.7 Dwellings 3.2 68.2 Industry 7.0 151.1 Manufacturing 11.9 254.8 Services 37.3 800.8 Total 100.0 2,145.5 Source: SAM (Chapter 8 of this book). Growth Analysis 149 industrial prices have been increasing consistently. Agriculture, which grew at a slower pace than the rest of the economy during the 1990s, has absorbed a surplus of (unskilled) labor released in the process of economic transformation. The fastest growing sector in the econ- omy--construction--has made a significant contribution to job crea- tion: the share of the labor force employed in construction has risen from 4 to 9 percent over the past five years. Job creation in industry in recent years was mostly concentrated in new private enterprises. Out of 26,500 jobs created in industry during 1999­2003, 22,000 were established in newly created private enterprises. Investments The sectoral composition of investments shaped the sectoral growth pattern. Figure 7.2 presents the structure of investments by sectors with housing representing about 30 percent of total investments dur- ing 1999­2003. Infrastructure has also attracted a considerable share of Figure 7.2 Sectoral Shares in GDP and Investments (average) 35 30 25 20 cent per 15 10 5 0 industry agriculture transport trade housing other sector share in investments 1999­2003 share in GDP 1999­2004 Source: WDI. 150 Part II: Detailed Analysis investment. Several profitable and competitive traditional sectors have attracted private investment during the process of economic transfor- mation (for example, gem cutting, brandy making, electric motors, hotels, high-technology manufacturing, and software services). A significant source of investment financing has been foreign direct investment (FDI). The major portion of FDI, both in infrastructure (tel- ecommunications through the privatization of the national telecom- munications company, ArmenTel) and in industry (privatization of the gas distribution network), was the direct result of large-scale privati- zation. Two Russian companies, including Gasprom and the energy corporation Itera, and the Greek OTE, are among the main investors in the country. The third largest foreign investor is Canada, with about 11 percent of the total investment, followed by the United States with 10 percent of the total cumulative investment. France, Luxembourg, the United Kingdom, Cyprus, Italy, Belgium, and Switzerland are other large investors in the country. FDI in the agro-processing sector, especially in beverages, has in several cases stimulated technology transfer and changes in the pro- duction pattern. One of the largest privatization deals in this sector involved the sale of the Yerevan Brandy Factory (YBF). The privati- zation of YBF generated strong linkages between farmers and buyers of agricultural produce. But this example is one of very few and an indication of the barriers that deter stronger private investment, both domestic and foreign, to the sectors with high export and growth, high value-added activities, and therefore high profit potential. Exports Further analysis of the investment structure is limited by the lack of detailed and consistent data on a micro level. The analysis of export performance gives a good indication of the sectors that have managed to attract investments. Armenia's exports have experienced impressive growth since the mid-1990s. In real terms, exports have grown by 13 percent per year since 1995. The share of exports in GDP has steadily grown and in 2003 the share of exports in GDP was 32 percent. Since 1995, Armenia has experienced impressive growth in its volume of trade--a trend only temporarily dampened by the Russia crisis. Table 7.2 shows the sources and sectoral distribution of exports in Armenia as given in 2002. The composition of exports has registered some encouraging devel- opments, with exports diversifying toward high value-added man- ufacturing goods and services. However, exports still remain quite concentrated, with precious stones and metals accounting for more Growth Analysis 151 Table 7.2 Sources of Exports in Armenia, 2002, by Sector Share in total exports Contribution to exports Sector (percent) (US$ million) Diamonds 37.0 258.0 Other 31.2 217.6 Heavy metals & machinery 14.4 100.4 Manufacturing 13.9 97.1 Electricity supply and dist. 1.9 13.4 Agriculture 1.6 11.5 Total 100.0 698.0 Source: SAM, 2002 (Chapter 8 of this book). than 40 percent of total exports. Uncut gems are imported into the country, are cut and sometimes mounted as finished jewelry, and then re-exported to Europe or the United States. Other large exports are food products and machine products. Armenia also demonstrates a high share of services exports. However, a closer look at the composi- tion shows that service exports are sensitive to transportation costs. Armenia's exports tend to be concentrated in the sectors with low or no transport costs (cut diamonds, services, electricity). The analysis highlights the following features of Armenian trade: The integration of Armenia into world trade remains limited and has not seen major improvements outside of the diamond trade. Armenian exports have become increasingly concentrated, even compared to other countries with a large share of mineral exports, which makes the exports more volatile and vulnerable to changes in external conditions, such as input arrangements and prices. Over the past decade the number of goods in which Armenia has demonstrated a revealed comparative advantage (RCA) has not expanded significantly (Table 7.3). Armenia also demonstrates a high share of services exports. However, a closer look at the composition shows that service exports are sensitive to transportation costs. Armenia's exports tend to concentrate in sectors with low or no transport costs (cut diamonds, services, electricity). Trade in services may have increased substantially, but trade data will not capture this, owing to transfer pricing by firms and the general bias toward "goods" trade by statistics departments (Figure 7.3). 152 Part II: Detailed Analysis Table 7.3 Armenia: Revealed Comparative Advantage, 2003 Revealed comparative advantage 2003 Diamonds, not industrial, not set 75.7 Beverages 16.1 Metalliferous ores and metal scrap 12.0 Electric energy 10.1 Nonferrous metals 6.6 Nonferrous metals 1.5 Fruits and vegetables 1.1 Iron and steel 1.0 Electrical machinery, apparatus 0.1 Source: World Integrated Trade Solution (WITS). http://wits.worldbank.org. Note: The index for country i good j is RCAij = (Xij /Xit)/(Xwj /Xwt)*100 where w = world and t = total for all goods. The RCA index compares the composition of exports of a country with the composition of total world exports. An RCA index higher than 1 indicates that a country has a strong RCA in the export of the product. An RCA index lower than 1 indicates that the country has a comparative disadvantage in that product. An RCA index equal to 1 indicates that the country has neither an advantage nor a disadvantage. LINKS BETWEEN GROWTH, EMPLOYMENT, AND POVERTY Impact of Growth on Poverty A strong and almost decade-long rise in national income and, to a lesser extent, a reduction in inequality have been the two factors that have contributed to poverty reduction in Armenia. Since 1994, Arme- nia has grown at a remarkable average annual rate of over 7 percent and by 2004 its real GDP reached the pre-transition level. In addi- tion to output growth, private foreign transfers--invested mostly in construction--and remittances have played a key role in raising the national income. Both sources of growth in aggregate demand have triggered a high induced household income multiplier2 and produced strong spillover effects that have benefited multiple households. The poorest population in Armenia saw the strongest increase in consumption. Estimates based on the Armenia Integrated Living Con- ditions Survey (ILCS) indicate that Armenia saw a considerable decline in extreme poverty during 1998­2004. About 17 percent of the population, or more than half a million people, have moved out of poverty in the past five years. About 9 percent of the population, or 0.23 million people, escaped the basic food deprivation. This evidence is consistent with the evolution of inequality for the period--the Gini coefficient Figure 7.3 Matrix: Classification of Armenian Exports by Composition and Destination, 2003 SICT 6672 68 11 28 67 89 35 72 05 Export Diamonds Nonferrous Beverages Metal ores Iron & Miscellaneous Electric Electrical Fruit & metals & scrap steel manufacturing energy machinery vegetables Israel Belgium Russian Federation Germany Netherlands UAE United States United Kingdom Switzerland Georgia Iran Ukraine Gr Greece owth China Turkmenistan Analysis Thailand Canada Belarus Source: WITS database. http://wits.worldbank.org. 153 Note: SICT: Standard International Trade Classification 154 Part II: Detailed Analysis decreased from a value of 0.299 in 1998­99 to a value of 0.259 in 2004--and indicates that the growth was accompanied by a compres- sion in the overall structure of consumption distribution. Growth has been channeled to the households through multiple sources that have increased both factor and nonfactor income. A strong growth in factor income from both labor and capital was accompa- nied by a stable growth in remittances and a consistent improvement in public transfer programs. In fact, household income from all main sources rose in real terms between 1998­99 and 2004. Factor income remained a dominant source of household income for all quintiles, followed by income from remittances and public transfers. Such a mul- tiplicity of income sources has benefited various types of households and resulted in a broad-based poverty reduction in Armenia. Growth in factor income mirrored growth in output3 and was based on large productivity gains that were due to structural changes and labor shedding in the process of labor rationalization, and to a lesser extent, to a rise in employment. Survey data point to a 17 percent unemployment rate, which is an estimated 35 percent reduction in the unemployment rate estimated in 1998­99, but it is still the high- est unemployment rate registered among the transition countries. In addition, much of the work force is to be found in low-paid subsist- ence agriculture, trade, the informal economy, and temporary employ- ment. While the demand for additional formal labor grew slowly, pro- ductivity gains secured stronger earnings for current workers with stable employment arrangements, with real wages seeing a 50 percent increase between 1998­99 and 2003. The investment contribution to growth was also considerable and the share of capital value added in Armenia's GDP has increased over time. Income from capital has been growing rapidly, benefiting those who owned and/or received the assets at the outset and/or are in the process of economic transfor- mation. Armenia's recent growth has also benefited from large foreign capital transfers invested mostly in the construction sector, which is characterized by a strong income multiplier effect on employment, and therefore on household incomes. In addition to factor income, both public and private transfers have played an important role in boosting household income in Armenia. The poor rely heavily on government transfers (such as pensions, social assistance, and other transfers), with this source representing about 20 percent of household income and the third largest source of income (after labor income and remittances) in the poorest quintiles. Social protection benefits have increased in the past five years in real per capita terms. Benefits have also improved in terms of targeting the poorest, with the family poverty benefit system becoming an effective mechanism for reaching the poor. Growth Analysis 155 Box 7.1 Reconciliation of National Accounts and the Household Survey Data The National Accounts succeed in understanding Armenia's underlying production technology, trade position, and final demand structure. How- ever, the use of a "representative" household agent for total demand makes it impossible to determine whether economic growth in Arme- nia is helping to alleviate poverty. In our study we have scaled total household incomes to meet total factor endowments and total (official) transfers from abroad, from the government, and from investment. For poor households, this leads to a significant divergence from reported income from the 2002 Household Survey. The most onerous problem when using household data is to rec- oncile household consumption against household income. Several of the poorer deciles in Armenia report far lower incomes than they report expenditures. Even when government and family transfers are included, expenditure for the poorest decile in Armenia is at least 10 times as large as income. Indeed, according to the Statistical Handbook based upon the 2002 Armenian Household Survey (Table 1, p. 51), 2002 total monthly per capita income was 12,776 drams. Of this, monetary income was 9,781 drams and nonmonetary (that is, own-produced value of consump- tion) monthly income was 2,995 drams. Annual monetary income, based on the survey, is then 117,372 drams per year (12 × 9,781), or US$204.83 per year, taken at market exchange rates. This implies that average income is US$0.56 per day. However, when using the National Accounts, we find that total value added (GDP at producer prices) is US$2.145 billion. Using 3.1 million as the national population, per capita income is US$1.90, more than three times as large as income from the Household Survey. During the reconciliation, some of the differences between rural and urban incomes can be attributed entirely to the investment account. Source: SAM (Chapter 8 of this book). Remittances and other private transfers by far exceed publicly pro- vided resources in Armenia.4 They accounted for more than 10 percent of GDP during this period and boosted consumption levels, includ- ing among the poor, thereby helping to reduce poverty. An increas- ingly large number of households receive remittances reaching 50 percent of the monetary income of recipient households. Remittances have a direct effect on the recipient households' well-being as well as strong secondary multiplier effects for those who do not receive 156 Part II: Detailed Analysis remittances directly. External remittances comprise a large share of overall monthly income for households that receive them, which sug- gests that remittances have significantly improved the well-being of the poorest households and have been critical for keeping these quin- tiles out of poverty. Impact of Growth on Employment Armenia's labor force comprises about 50 percent of the total pop- ulation, which is comparable with the situation in other European countries. The situation with employment, however, is much worse in Armenia than in comparator countries. In contrast to the strong growth performance in Armenia, the evolution of aggregate employ- ment during the 1990s was negative. The labor force remains poorly utilized; the labor participation rate is low and unemployment is ram- pant (over 19 percent). The formal employment level is low and the main source of employment growth until recently was through the expansion of self-employment and low value-added agriculture and trade. High emigration has been negatively affecting the domestic supply5 of young and skilled labor, combined with slow growth on the demand side especially with slow growth in skilled jobs. Data on wages by professions suggest that the strongest increase was recorded in skilled labor, but, as mentioned above, the number of skilled job positions created was limited. The upswing has been shared by a range of professions, including unskilled and semiskilled labor. An economy's long-term growth is driven by the evolution of its underlying supply factors of production and physical and human cap- ital, and by the development of total factor productivity. In Armenia, the positive output trend has not translated into employment growth, and formal employment is currently below that reported in the early 1990s (Figure 7.4). Total factor productivity growth seems to be cap- turing most of the growth pattern, reflecting efficiency gains stimu- lated by structural changes over the mid-1990s and early 2000s. As with other transition countries, the surge in unemployment in Armenia was anticipated given the massive economic and geographic reallocation of labor during the transformation. This phenomenon, called "jobless growth," is not specific to Armenia but is typical of most transition economies, where initial growth is mostly based on large productivity gains due to structural changes and labor shedding in the process of labor rationalization and more efficient utilization of labor resources, rather than through an increased use of the labor force. In the process of a country's search for new equilibrium, the destruction of unproductive jobs is accompanied by the creation of new and more productive jobs. However, job losses usually exceed job gains, leading to a net fall in employment. Growth Analysis 157 Figure 7.4 Real Growth and Factors of Production, 1997­2004 15 1,000 800 10 600 rate 5 million growth 400 US$ 0 200 -5 0 1997 1998 1999 2000 2001 2002 2003 2004 year employment growth rate (adjusted) (left-hand side) real GDP growth (left-hand side) capital investment (US$) (right-hand side) Source: Armenian Department of Statistics. According to the household surveys, which are based on self-report- ing, unemployment overall was 29 percent in 2002, down only 2 per- centage points from 2001, with most of the decline being in Yerevan, where unemployment fell by 7 percentage points between 2001 and 2002. The persistence of high unemployment and underemployment even after more than a decade of economic growth may reflect some underlying structural problems. The pace at which the Armenian economy will be moving toward its new equilibrium, and the pattern it will assume, will depend on a variety of factors. We focus our analysis on a few of these factors which are relevant to our analysis (for example, the issues related to the flexibility of labor markets are deliberately omitted as they are being examined in the paper on labor markets). Strong labor productivity growth and an increase in the unit labor cost of current workers have crowded out the short-term potential for new employment. In general, growth in labor productivity, if associ- ated with slower growth in real wages, reduces unit labor costs and 158 Part II: Detailed Analysis creates room for new employment. In Armenia, rapid growth in labor productivity was accompanied by an increase in hourly compensation above the productivity growth level. Armenia started from a low level of productivity but exhibited one of the fastest productivity growth rates among the comparator coun- tries. Labor productivity in Armenia grew continuously over the last decade, and this growth accelerated during the early 2000s (Figure 7.5). During this period Armenia realized much of the "catch up" potential which arose following the economic transformation. As a result of this rapid productivity growth, the productivity gap between Armenia and industrialized countries has narrowed; how- ever it remains substantial and accentuates the importance of sustain- ing recent growth rates. The high growth rate of labor productivity was mainly due to Armenia's comparative advantage of an educated and skilled workforce and the country's strong tradition of having highly skilled craftsmen. Figure 7.5 Real Wage Growth and Productivity Growth, Unit Labor Cost (US$), 1998­2004 25 0.5 20 0.4 15 0.3 rate US$ growth 10 0.2 5 0.1 0 0 1998 1999 2000 2001 2002 2003 2004 year real wage growth (left-hand side) productivity growth (left-hand side) unit labor cost (right-hand side) Source: Armenian Department of Statistics. Growth Analysis 159 Labor productivity growth in Armenia was outperformed by the increase in labor compensation. The unit labor cost (ULC) (influenced by diverse factors such as hourly compensation, productivity, and the exchange rate) gradually increased to about 46 percent of the U.S. ULC level (about 44 percent of the euro-area level) from as low as less than 3 percent in the early 1990s, and secured solid earnings for current workers.6 However, the benefits of recent economic growth have not generated higher employment and the increase in the labor cost has partly contributed to jobless growth (Figure 7.6). Figure 7.6 illustrates the mismatch between the rapidly expand- ing sectors and labor productivity. The contrast between the rapidly growing sectors and formal employment indicates that the majority Figure 7.6 Labor Productivity by Sector 3.0 2.5 2.0 1.5 index 1.0 0.5 0.0 e services uction industry average agricultur economic constr manufacturing sector 4.5 4.0 3.5 3.0 2.5 index2.0 1.5 1.0 0.5 0.0 & food products uction power stonesprocessing average mining banking quarrying economic constr mineralecious transportation communication pr & and sector Source: Armenian Department of Statistics. 160 Part II: Detailed Analysis of the labor force has benefited only modestly from rapid economic expansion. Growth in Armenia has been concentrated in non-tradables and few tradable sub-sectors. Contrary to the industrialization pattern of the East Asian Tigers (like the Republic of Korea and Taiwan, China) with highly diversified export structures, recent growth in Armenia has been characterized by a mix of export-orientation toward capital intensive goods (such as scrap metal), commodities with low trans- portation costs (such as diamonds), and exports with a high price/ volume ratio (such as brandy), accompanied by import substitution of more labor-intensive commodities (such as subsistence agriculture). The export bias toward capital-intensive goods is explained by the high costs of exporting goods (and services) from Armenia and there- fore the concentration of exports around services/commodities with a high ratio of price to volume. This particularly high cost structure is a combination of common "transition economy" factors and Armenia's specific factors: In Armenia, as in any other transition economy, new jobs cre- ated differ in their salient characteristics from old jobs that were destroyed, and employment growth relies on business start-ups. The number of new businesses and new jobs created is directly associated with the profit margins generated by the new firms, while the costs of the new firms are generally high owing to initial investment and working capital needs, marketing, skills mismatch, high risks, and uncertainties (see subsequent chap- ters for detailed analyses). At the same time, new firms are con- fronted by severe resource constraints. Savings are low, financial intermediation is expensive, stock markets are underdeveloped, and insurance services are virtually nonexistent. These high costs are inflated by additional charges related to corruption, which is also common in other transition economies but to varying degrees (other chapters in this report will try to benchmark Armenia relative to other transition economies). The high cost of doing business in Armenia and the high transaction costs push businesses into the informal economy, reduce invest- ment opportunities, and prevent Armenia's participation in the global division of labor based on production fragmentation and direct impact on the number of jobs created. In addition to the above, Armenia faces the challenge of a small domestic market and market isolation. This makes the coun- try dependent on exports and at the same time makes exports expensive because of high transportation costs. Growth Analysis 161 And finally, the high costs of backbone services (such as air fares and telecommunications), higher, in fact, than in the countries that do not face structural problems in market access, exacerbate the cost of doing business to levels not affordable for an average start-up with a modest profit margin. As a result, only exports that have very low transportation costs (such as diamonds and electricity exports to neighboring countries) or exports with a high price/volume ratio (such as brandy)--in other words, capital-intensive industries--have proved to be profitable in Armenia and have attracted investment. On the other hand, labor- intensive commodities such as light manufacturing, which requires bulk exports, have been developing much more slowly. In addition to the above, the sectors that had the opportunity to reduce costs by capitalizing on the existing export markets through a brand name (for example, brandy in Russia) or an established chain of production (such as precious stones refinement) have continued to follow the high cost of doing business. Another arrangement that allowed new businesses to operate was becoming informal and oper- ating underground. Growth Spillover Channels Armenia's growth has recently benefited from an increase in remittances and capital transfers, similar to the resource boom type of growth, where sources of growth rest on concentrated and volatile internal sources of demand financed from external transfers. Sources of growth rely heavily on both consumption driven by remittances and private transfers7 (dis- cussed earlier in the chapter), and on investment, mainly in construction, which relies heavily on foreign capital transfers. In general, the high share of value added in production leads to a greater contribution of that particular activity to GDP and growth. A strong reliance on intermediate inputs leads to low value added in production and is associated with low value added multipliers. Typically, sectors with high value added and intersectoral linkage multipliers would have a higher growth impact, but the gains from a large multiplier depend on the sector's size. Armenia's economy, especially its industrial sector's reliance on intermediate inputs, is quite high. Despite low savings rates and low effective taxes, which represent other leakages from output, the total value added as a share of total production is about 50 percent. The share of value added in industry is just 30 percent and the share of manufacturing only 16 percent; and the GDP multipliers of these sectors are lower than the 162 Part II: Detailed Analysis economy average (close to 1 as compared to about 1.4 on average in the economy). Machinery and jewelry cutting have the lowest share of value added in production, which implies that these industries are comprised mostly of intermediate inputs. In contrast, the share of value added is quite high in construction, agriculture (crops, vegeta- bles, and fruits), and services (utilities and banking), which implies large value-added multipliers (above 1.5) for these three sectors and the much stronger impact of these sectors on GDP growth than industry and manufacturing. Growth in these sectors is more likely to contribute to higher personal income for Armenians. Relatively, these sectors demonstrate high household multipliers. Over time, the share of capital value added in Armenia's GDP has increased, and although this has led to high overall incomes, it has had a relatively narrow impact on poverty reduction, as the earners of profits in a capital-biased, value-added economy depend on how these profits are distributed. Some industries, such as cellular telecom- munications, mining, or utilities may exhibit a high degree of value added in production, but since those industries are capital intensive, their contribution to personal income will depend on who collects the capital rents. For example, a foreign-owned mine may capture signifi- cant rents, but most of this income could be sent off-shore. Intersectoral linkage multipliers (Box 7.2) are highest in some serv- ices sectors, particularly in real estate, banking, and insurance. The intersectoral linkage multipliers are particularly low in some manufac- turing sectors, such as light manufacturing and machinery and equip- ment that rely on imported inputs (the share of imports in absorp- tion has recently increased to 45 percent). The mismatch between the level of GDP and intersectoral linkage multipliers and the share of the sector in value added indirectly point toward distortions in resource allocation, suggesting that there may be a disincentive for firms to invest in potentially highly profitable activities. While the largest intersectoral linkage multipliers in agriculture are in high value-added crops such as fruits and vegetables, the share of these sectors in Armenia's GDP is lower than the share of those with much lower multipliers. This suggests that some sectors with high growth potential have a much lower impact on growth than they potentially could have if they were to develop significantly in the future. Capital- izing on such high multiplier sectors would require finding ways of channeling investments into such sectors through a set of policy actions, including eliminating price distortions. To strengthen the backward and forward linkages from agriculture to agro-processing and manufactur- ing, the authorities should develop a strategy that enhances the efficient use of Armenia's agricultural base, including through improved access to markets and an increase in productivity. Growth Analysis 163 Box 7.2 Linkages in the Model Based on the Social Accounting Matrix The model based on the social accounting matrix (SAM) provides insights about the input and output relationships between goods and services markets, as well as the interlinkages between the demand for labor and capital in the process of production. This is made possible by analyzing a set of comparable direct and indirect intersectoral linkage multipliers in the context of the relative importance of each economic activity. Multipliers in the SAM-based model reflect how much a unit-increase in the demand and production of a commodity affects the demand for more commodities. The multiplier effect entails a higher demand for activities leading to higher wage incomes and profits (that is, higher household incomes), which create the demand for domestic or imported commodities. Direct multipliers: Value-added multipliers--show how much a unit-increase in the demand for a commodity directly affects the demand/ production of the same commodity. Intersectoral linkage multipliers--illustrate how much a unit- increase in the demand for a commodity increases the demand/ supply of commodities produced by other activities. Indirect multipliers: Induced household income multipliers--show how much a unit- increase in the demand for a commodity affects household income. Employment multipliers--show how many jobs are ultimately created directly or indirectly as a consequence of a unit-increase (1 million dram) in the demand for a commodity. The level of multipliers reflects the level of value added in produc- tion. Strong reliance on intermediate inputs leads to low value added (to labor and capital) and, hence, small multipliers. This could be the result of large leakages from an economy, due to a large proportion of imported inputs, taxes, and savings. Strong linkages entail direct and indirect spillovers of subsectoral growth on growth in other sectors, employment generation, and income distribution. Source: SAM (Chapter 8 of this book). 164 Part II: Detailed Analysis Remittances represent a considerable source of household income (US$207 million approximately)8 and are one of the main sources of foreign exchange for the country. An increasingly large number of households receive remittances in Armenia. Within these households, remittances tend to account for around 50 percent of their monetary income. The overwhelming part of remittances is spent on current consumption and only a small percentage is used for investment. Remittances have a direct effect on the recipient households' well- being as well as strong secondary multiplier effects. We have used the social accounting matrix to simulate the multiplier impact of remit- tances on the national economy. Impact on employment and household incomes by sector. Construction-- the fastest growing sector in Armenia--has had a strong impact on employment as well as on household incomes (see Table 7.4 for the employment and income multipliers as derived from an SAM con- structed for Armenia). Direct short-term spillover effects associated with the construction sector are strong, as is illustrated by large mul- tipliers and validated by the high share of the construction sector in the total economy's value-added and low-import content. The construction sector is strong in generating jobs, but the perma- nence of such jobs will depend on continued activity. The prospect of Millennium Challenge Account funding is, therefore, encouraging for employment. But, equally, a possible decline in construction would depress the direct demand for multiple economic activities. The spill- overs from these would be strong. The decline in incomes would result from a reduction in incomes generated in the construction sec- tor, as well as indirectly from other sectors. Suppressed household incomes would, in turn, depress the demand for commodities and services produced domestically. This would produce multiplier effects and exacerbate the effect of the original shock. There are a number of subsectors in agriculture, services, and manufacturing that are relatively stable and could potentially pro- duce stronger spillover effects and contribute significantly to growth, sustainable employment, and poverty reduction (the value-added share in production is high, and direct and indirect multipliers are quite significant). However, their ability to attract investments seems to be distorted. Their relatively low size and high export share in total production indicate that they may react strongly (may easily gener- ate double-digit growth and significant exports) when the constraints to growth are eased and/or incentives are provided. Capitalizing on such activities will be critical for increasing growth sustainability, diversifying the sources of exports, and creating better-paid jobs. Growth potential and the spillover effect of a number of activities that have not been growing rapidly are high in the sectors with the following characteristics: Growth Analysis 165 A high value-added share in production, particularly in some agriculture subsectors (above 60 percent) but also in services (banking and insurance and real estate). The share of value added is quite high in construction, agriculture (crops, vegetables, and fruits), and services (utilities and banking), which implies large value added multipliers (above 1.5) for these three sectors and a much stronger impact of these sectors on GDP growth than industry and manufacturing. Growth in these sectors is more likely to contribute to higher personal income for Armenians. Related to these characteristics, these sectors demonstrate high household multipliers. High intersectoral linkage multipliers such as banking and insur- ance and real estate. The large share of exports in total production for processing sec- tors and, at the same time, weak export performance of these activities indicate that: (i) the potential of these sectors is under- utilized and (ii) growth potential in these sectors can most likely be realized through export. The recent strong export expansion has had a limited impact on poverty reduction. Exports are concentrated in diamonds (37 percent of total exports, or US$ 258 million in 2002). The sector's contribution to the economy's value added is just 2 percent as the sector relies heavily on diamond imports (Table 7.5). This is an enclave activity with modest spillover effects for growth. In terms of employment, the sector does not create many jobs, but the jobs created are high-quality ones and are critical for the economy. In addition to diamond process- ing, export diversification in other semiskilled and skilled labor and higher value-added activities would generate stronger spillover effects and would translate into job creation and overall poverty reduction. The results of the simulations show that further diversification of the economy is critical for reducing the vulnerability of growth to external shocks and maintaining a reasonable growth momentum in the medium to long terms. As earlier sections of this chapter indi- cated, the economy has a small nondiamond production/export bas- ket. With the exception of a few traditional commodities and power, this basket does not reveal a comparative advantage in any particular commodity. However, it does suggest a natural comparative advan- tage through its linkages to the country's agricultural base and other economic resources as well as skilled human capital. The challenge for policy makers is to provide the general policy and institutional underpinnings for faster development of formal private sector activi- ties, which (i) tap into the higher-value elements of this basket that can create high-wage employment, and (ii) capitalize on the external demand by meeting the scale and quality requirements. 166 Table 7.4 Direct and Indirect Multipliers Linkage Induced Capital Part Share Own with other HH income Labor and land Sector in GDP VA/output multiplier sectors multipliers multiplier multiplier II: Detailed 1 grn: Wheat, potatoes, and legumes 4.6 63.5 1.1 4.9 1.6 0.8 0.8 2 vfr: Vegetables and fruits, including grapes and dried fruits 5.4 87.1 1.1 5.6 1.9 0.9 1.0 3 vol: Vegetable oils and fats 0.0 6.0 1.0 3.2 0.9 0.4 0.5 Analysis 4 ocr: Crops not elsewhere classified 3.8 78.7 1.1 5.5 1.8 0.8 1.0 5 mil: Dairy products, including eggs and milk 7.8 49.1 1.4 5.0 1.6 0.8 0.8 6 omt: Beef, pork, mutton, and poultry meat 3.1 42.0 1.1 5.1 1.6 0.8 0.8 7 enr: Energy--oil and natural gas 0.0 0.0 1.0 0.0 0.0 0.0 0.0 8 min: Mining and quarrying 1.8 59.8 1.2 5.4 1.7 0.7 1.0 9 fod: Food processing and beverages 7.4 36.7 1.4 4.6 1.5 0.8 0.7 10 tbc: Tobacco products 0.6 17.0 1.1 3.2 0.9 0.4 0.5 11 lmf: Light manufacturing and textiles 0.2 10.1 1.2 2.8 0.8 0.4 0.4 12 crp: Chemicals, rubbers, and plastics 0.5 19.6 1.1 3.5 1.0 0.5 0.5 13 mnm: Mineral products and precious stones 1.9 15.1 1.0 2.9 0.8 0.4 0.4 14 mtl: Metals and metal products 2.0 39.7 1.1 4.5 1.4 0.7 0.7 15 mch: Machinery equipment and motor vehicles and precision optical equipment 0.3 6.5 1.1 2.3 0.6 0.3 0.3 16 omf: Other manufacturing 0.8 29.3 1.1 3.5 1.0 0.5 0.6 17 utl: Electricity, gas, and water supply 5.2 73.5 1.3 5.4 1.8 0.7 1.1 18 ele: Electricity supply and distribution 4.8 62.6 1.2 4.5 1.5 0.7 0.8 19 con: Construction 13.9 57.6 1.1 4.7 1.5 0.7 0.9 20 trn: Transport and communications 7.1 58.9 1.2 5.0 1.6 0.8 0.8 Table 7.4 (continued) Linkage Induced Capital Share Own with other HH income Labor and land Sector in GDP VA/output multiplier sectors multipliers multiplier multiplier 21 trd: Retail and wholesale trade and public catering 13.5 64.5 1.3 5.0 1.7 0.7 1.0 22 bnk: Banking, lending, and insurance 1.7 69.1 1.1 5.6 1.8 1.0 0.8 23 gov: Governance, defense, and public procurements 9.2 54.5 1.2 5.2 1.7 1.0 0.7 24 osr: Other services not elsewhere classified 1.1 25.2 1.1 4.2 1.2 0.5 0.7 25 dwe: Housing and dwellings 3.2 73.6 1.1 5.9 1.9 0.5 1.4 Total 100.0 n.a. n.a. n.a. n.a. n.a. n.a. Total agriculture 24.8 58.4 1.1 4.9 1.6 0.7 0.8 Total manufacturing 11.9 29.0 1.1 3.5 1.0 0.5 0.5 Gr Total industry 8.9 15.9 1.1 3.4 1.1 0.5 0.6 owth Total construction 13.9 65.7 1.1 4.7 1.5 0.7 0.9 Total services 40.5 55.5 1.2 5.1 1.6 0.7 0.9 Analysis Total 100.0 49.5 n.a. n.a. n.a. n.a. n.a. Source: SAM, 2002 (Chapter 8 of this book). Note: Here, share of VA in output means that for every dram of production only x percent is value added. So, while production or output from an activity can be very large, the value added can be very small because of the high share of intermediate, especially imported, inputs. 167 168 Table 7.5 Armenian Economy, Sectoral Structure Depicted by SAM Part Share Share Share II: Sector in GDP VA/Output in exports in nondiamond exports Detailed grn: Wheat, potatoes, and legumes 4.6 63.5 0.0 0.1 vfr: Vegetables and fruits, including grapes and dried fruits 5.4 87.1 0.8 1.3 vol: Vegetable oils and fats 0.0 6.0 0.0 0.0 Analysis ocr: Crops not elsewhere classified 3.8 78.7 0.7 1.1 mil: Dairy products, including eggs and milk 7.8 49.1 0.1 0.1 omt: Beef, pork, mutton, and poultry meat 3.1 42.0 0.0 0.0 enr: Energy--oil and natural gas 0.0 0.0 0.0 0.0 min: Mining and quarrying 1.8 59.8 5.5 8.8 fod: Food processing and beverages 7.4 36.7 7.3 11.6 tbc: Tobacco products 0.6 17.0 0.5 0.8 lmf: Light manufacturing and textiles 0.2 10.1 4.2 6.6 crp: Chemicals, rubbers, and plastics 0.5 19.6 1.2 1.9 mnm: Mineral products and precious stones 1.9 15.1 37.0 mtl: Metals and metal products 2.0 39.7 6.4 10.1 mch: Machinery equipment and motor vehicles and precision optical equipment 0.3 6.5 8.0 12.7 omf: Other manufacturing 0.8 29.3 0.7 1.1 utl: Electricity, gas, and water supply 5.2 73.5 0.0 0.0 ele: Electricity supply and distribution 4.8 62.6 1.9 3.0 Table 7.5 (continued) Share Share Share Sector in GDP VA/Output in exports in nondiamond exports con: Construction 13.9 57.6 0.9 1.4 trn: Transport and communications 7.1 58.9 12.6 19.9 trd: Retail and wholesale trade and public catering 13.5 64.5 0.0 0.0 bnk: Banking, lending, and insurance 1.7 69.1 1.0 1.5 gov: Governance, defense, and public procurements 9.2 54.5 1.1 1.7 osr: Other services not elsewhere classified 1.1 25.2 10.1 16.1 dwe: Housing and dwellings 3.2 73.6 0.0 0.0 Total 100.0 n..a. 100.0 63.0 Total agriculture 24.8 58.4 1.6 2.6 Total manufacturing 11.9 29.0 28.3 44.9 Total industry 8.9 15.9 42.5 8.8 Gr Total construction 13.9 65.7 0.9 1.4 owth Total services 40.5 55.5 26.7 42.3 Total 100.0 49.5 100.0 100.0 Analysis Source: SAM (Chapter 8 of this book). Note: Here, share of VA in output means that for every 1 dram of production only x percent is value added. So while production or output from an activity can be very large, the value added can be very small because of the high share of intermediate, especially imported, inputs. 169 170 Part II: Detailed Analysis The impact of growth on formal job creation was limited, as the sec- tors that have experienced strong growth in Armenia have not been labor-intensive (Table 7.6). On the contrary, agriculture, which has been growing more slowly than the rest of the economy, has absorbed surplus labor released in the process of economic transformation. The government--the most labor-intensive activity in Armenia--has been negatively affected by fiscal adjustment. Other labor-intensive sectors (such as banking and light manufacturing) have been growing slowly. The main reason for the unfavorable progress in the light manufac- turing sector is related to high transportation costs, such as aviation. The construction sector, which has been a leading growth sector, has provided quite a few employment opportunities. However, the sus- tainability of these jobs depends on the sustainability of the source of its financing, which is volatile. Diamond cutting, which is a sector with good jobs, is capital-intensive but the contribution of labor value added in the sector is not. The effective reallocation of jobs released as a result of the termination of large construction projects depends on the ability of the economy to provide alternative opportunities outside of the construction sector in case of demand contraction. The causes of the centering of growth on sectors that are not labor-intensive are explored in Chapter 11 of this book. The mismatch between rapidly expanding sectors and labor productivity. The contrast between the rapidly growing sectors and formal employ- ment indicates that the majority of the labor force has benefited only modestly from rapid economic expansion. Armenia has shown some signs of export diversification, but a number of sectors with high growth potential remain idle. The mis- match between the level of GDP and intersectoral linkage multipliers and the share of the sector in value added indirectly point toward the distortions in resource allocation, suggesting that there may be a disin- Table 7.6 Sources of Employment in Armenia: Distribution by Sector, 2002 Share in total employment Contribution to employment Activity (percent) US$ thousands Agriculture 49.5 548 Construction 4.8 53 Industry 1.6 18 Diamonds 1.4 15 Manufacturing 10.6 117 Services 32.1 355 Total 100.0 1,106 Source: SAM, 2002 (Chapter 8 of this book). Growth Analysis 171 centive for the firms to invest in potentially highly profitable activities. Earlier sections have noted that despite the apparent diversification of the nongold export basket, very few--mostly primary--exports reveal a comparative advantage and even fewer export items show signs of emerging competitiveness in regional markets (the dairy indus- try and heavy manufacturing are two examples). The present export basket, however, reflects strong backward linkages to the country's natural resource endowments and agricultural base and other eco- nomic resources, as well as skilled human capital. As reflected in the presence of supply chains in several sectors, there are also clear for- ward linkages, especially with respect to agro-processing and light manufacturing, although these linkages are weak, particularly toward the upper end of the supply chain. A diversified export basket that takes advantage of the presence of these supply chains will not only facilitate the production and export of raw materials from agriculture and industry, but will also facilitate higher value-added processed and manufactured products from the same supply chains (such as canned vegetables and fruit juices). High value-added agricultural products. The trade data analysis sug- gests that crops such as vegetables and fruits, which presently have a relatively high share of GDP but a miniscule share of nondiamond exports, have significant growth potential--evident from their high value-added shares in total production (over 80 percent)--and export potential. The existing forward linkages of these commodities to agro- processing suggest that if their supply is sufficiently expanded, they can generate even higher value in the form of processed food. To strengthen the backward and forward linkages from agriculture to agro-processing and manufacturing, the authorities should develop a strategy that enhances the efficient use of Armenia's agricultural base, including through improved access to markets. Manufacturing (particularly textiles, chemical production, machinery and equipment, and motor vehicle production) can aid export diversification. The large share of exports in total production in the processing sectors, together with the observed weak export performance of these activi- ties, indicate that (i) the potential of these sectors is underutilized, and (ii) growth potential in these sectors can most likely be realized through export. High transportation costs and domestic barriers to the development of the higher end of these supply chains are the main obstacles to the expansion of these subsectors (for more details, see Part II, Chapters 9­11). Their relatively low size and high export share in total production indicate that they may react strongly (eas- ily generating double digit growth and significant exports) when the constraints to growth (such as policy-induced high business trans- actions costs or aviation and telecommunications costs) are eased. Capitalizing on such activities will be critical for increasing growth 172 Part II: Detailed Analysis sustainability, diversifying sources of exports, and creating better paid jobs. The main distortion for a stronger interest in high value-added activities with large export potential is transportation costs. Armenia has experienced a sharp drop in its trade transported by air in recent years, despite its landlocked nature and its subjection to a blockade by Turkey and Azerbaijan. The cost of air cargo in Armenia is much higher than that in other landlocked countries such as Kazakhstan, the Kyrgyz Republic, or Mongolia, or the semi-landlocked Turkmenistan. Such high costs have caused a relative shift toward road transport (despite poor infrastructure) and even railways. Liberalizing air trans- port by expanding the participation of several carriers could greatly enhance export competitiveness in Armenia. Finally, a number of subsectors in services such as banking and insurance, ITC, and tourism present cases where the potential has not been fully exploited. In addition, these sectors demonstrate high intersectoral linkage multipliers, which suggests that these sectors potentially produce stronger spillover effects and contribute more to growth, sustainable employment, and poverty reduction (the value- added share in production is high and direct and indirect multipli- ers are quite significant). However, their ability to attract investments seems to be distorted. A key barrier, especially for the services sec- tor, is the lack of competition and the high prices of telecommunica- tions. The telecommunications infrastructure in Armenia is currently one of the least developed in the CIS, despite being one of the first in the CIS to privatize. The 15-year monopoly granted to the single private investor since the privatization of ArmenTel in 1997 to the Greek operator OTE, coupled with the government's ineffective regu- lation of the operator, has led to severe underinvestments and has stifled growth even in mobile phones, where the penetration ratio in Armenia is one-tenth that of its neighbors. The extremely high cost and poor quality of telecommunication services has handicapped the country in production and exports. ANALYZING EXTERNAL SHOCKS AND SOURCES OF VULNERABILITY IN ARMENIA: SAM-BASED SIMULATIONS OF ALTERNATIVE SCENARIOS Assumptions and Scenarios As discussed in the previous sections, Armenia has demonstrated a strong growth performance in the last several years. The generation of baseline growth does not seem to be a constraint for Armenia. The challenges that Armenia faces seem to be (i) how to ensure growth Growth Analysis 173 sustainability, and (ii) how to enhance a growth pattern that secures job creation and rising household incomes. Based on these considerations and taking into account the previous analysis of the growth pattern, we have decided to focus our simula- tions on the following issues: Construction is the leading, yet unstable, growth sector that strongly relies on a highly volatile source of financing--capital transfers. Direct short-term spillover effects associated with the construction sector (illustrated by large multipliers and validated by the high share of the construction sector in the total econo- my's value added) are strong. The sector may not be generating long-term sustainable employment, nor may it be a future source of growth, but the first-round effects on growth and employment seem to be large. A negative shock in the sector could therefore trigger a decline in growth, a reduction in employment, and an increase in poverty. The sector is not a viable source of growth unless accompanied by growth in export-oriented, high value- added activities that generate future sources of income through a demand for skilled jobs. An economy cannot grow by sim- ply continuing to build houses and infrastructure financed by uncertain external flows, unless it ensures that the demand for the new houses and infrastructure is strong and the returns on investment are generated. Exports are quite volatile, with diamonds representing the most important source of revenue from exports (37 percent of total exports, or US$258 million in 2002). Diamonds are susceptible to price fluctuations as well as conditions in inputs markets and are therefore vulnerable to external shocks; the sector's share in GDP, however, is only 2 percent. Although the sector is also an important source of foreign exchange, it is an enclave activity with modest spillover effects for growth in terms of employ- ment. The sector does not create many jobs and the jobs created are highly skilled. In this regard, export diversification toward high value-added activities should be pursued to reduce export concentration and export reliance on this highly volatile source of foreign exchange and to safeguard the economy from the negative impact of external shocks. As was mentioned above, remittances are an important source of household incomes (about US$207 million). Remittances are, however, an unreliable source of income, and are perhaps unproductive if they are channeled toward consumption alone, as may be occurring in Armenia. While this is an unsustain- able source of growth, as it is consumption-based and does not 174 Part II: Detailed Analysis generate long-lasting income and jobs, its first-round effects are large. Therefore, the negative shock associated with the reduc- tion of remittances may have a devastating effect on the econ- omy and on household incomes. There is a need to build an incentive mechanism to channel a portion of these remittances into savings that can be invested in sectors with future growth potential. Simulation Results The results of the simulations of a negative shock on cut-diamond production are provided below. We have assumed a 50 percent decline in diamond exports (that is, about US$135 million). The effects of this export decline on growth, employment, and poverty are relatively moderate due to the enclave character of the diamond activity (the sector's contribution to GDP is about 2 percent) and its strong reli- ance on unprocessed diamond imports. The diamond sector's link- ages with the rest of the economy are not as strong as they would appear, judged by the large share of diamonds in exports. The effect of a comparable shock on construction would be much stronger, as the sector has large multipliers and a relatively low import share. The impact is also stronger on employment. This has an important implication for growth impetus in Armenia, which currently relies strongly on construction. This is a sector fueled by investments--the most volatile component of the aggregate demand. In 2004, as a result of the termination of the Lyndsey project, with the associated US$100 million in injections into the economy, the construction contribution to GDP declined substantially and many people employed in the sector were laid off. A decline in construction would depress the direct demand for multiple economic activities, with strong spillovers to the rest of the economy. There would be a decline in incomes resulting from lowered income in the construction sector as well as, indirectly, in other sectors. Suppressed household incomes would, in turn, lower the demand for commodities and serv- ices produced domestically and, through multiplier effects, would exacerbate the effect of the original shock. The third simulation was conducted to see the effect of a com- parable shock to remittances. The associated negative effect on GDP was stronger than in the diamond export shock case but weaker than in the construction case owing to smaller value added multipliers. However, this shock had the strongest impact on household incomes (­11 percent). Based on our analysis, we would argue that construction is a good but volatile source of growth and, in order to cushion the economy against its sudden decline, the economic development policy should Growth Analysis 175 enhance more stable sources of growth and employment in the ser- vices and manufacturing sectors. The manufacturing sector has lower multiplier effects in the short run but provides sustainable employ- ment with a strong potential for future income generation. We have simulated a case when the sectors with high export poten- tial are given incentives, or when certain barriers to trade are elimi- nated (for example, transportation costs are reduced). The results of the simulations show that this scenario has a strong positive effect on growth, household incomes, and job creation.9 This positive effect on employment is even stronger when we take into consideration that jobs that are created are sustainable and well paid (Table 7.7). Future Growth Prospects and Desirable Growth Patterns with Stronger Spillover Effects The intersectoral linkages and transmission mechanism, which in recent years enabled strong economic growth to penetrate down to households through higher incomes, are likely to prevail in the medium term. Stronger growth in export-oriented sectors, driven by a shift toward higher value­added production, would clearly bring higher overall GDP growth in the medium term. Such a shift would also be instrumental in preparing the economy to face the difficult challenge of maintaining modest growth rates should any of the shocks described above materialize. For purely illustrative purposes we have also demonstrated the result of a 10 percent shock to agriculture output through higher pro- ductivity. However, higher productivity may be commensurate with lower prices, especially if the agricultural goods cannot be exported. In September 2004, there was a good melon crop and all of the farm- ers had excess melons. As there is only so much demand for melons on the domestic market, many of the melons were left to rot, since no one would buy them, even at near-zero prices. Unfortunately, water- melons are low-value per pound, so the melon growers are not able to sell them internationally. This underscores the importance of linkages of agriculture to agro-processing, which may boost productivity and create additional demand. Model Simulations: Sustainable Growth and Job Creation, Using Export Diversification The results of the simulations show that further diversification of the Armenian economy is critical to reduce growth vulnerability to external shocks and maintain a reasonable growth momentum in the medium to long terms. As earlier sections of this chapter have indicated, the Armenian economy presently has a relatively small 176 Part II: Table 7.7 Tables Generated Using the 2002 SAM-based Model for Armenia Detailed Effect on labor Effect on jobs income, US$ & numbers & percent change percent change Analysis Effect on GDP, Effect on household income, Shock US$ & percent change Skilled Unskilled Skilled Unskilled US$ & percent change I. Negative shocks Diamonds (­50% = US $135 m) ­109.5 ­8.6 ­44.8 ­5,500 ­52,700 ­109.5 (­5.1%) (­3.7%) (­5.5%) (­3.8%) (­5.5%) (­4.4%) Construction (­30% = US$135 m) ­204 ­13 ­76 ­8,300 ­89,500 ­204 (­9.5%) (­5.7%) (­9.3%) (­5.8%) (­9.3%) (­8.3%) Remittances (­5.5% = US$135 m) ­139.2 ­10.1 ­64.1 ­6,400 ­63,600 ­275 (­6.5%) (­4.4%) (­6.6%) (­4.4%) (­6.6%) (­11.1%) Source: SAM (Chapter 8 of this book). m = million. Growth Analysis 177 nondiamond production/export basket. With the exception of a few traditional commodities and power, this basket does not reveal a com- parative advantage in any particular commodity. However, it does suggest a natural comparative advantage through its linkages to the country's agricultural base and other economic resources as well as skilled human capital. CONCLUSIONS The single most important factor explaining the recent poverty decline in Armenia was high growth combined with declining inequality. The principal sources of growth, household income, and foreign exchange in Armenia (that is, construction, remittances, and con- centrated exports), are quite concentrated, although there are clearly some elements of economic and export diversification. Trade and export-oriented growth beyond diamond processing should become a leading strategy of the development model pursued by the gov- ernment. While domestic savings need to continue to grow, FDI as well as portfolio investment as opposed to private transfers should become more important drivers of economic growth. More generally, the country should open up to benefits from intensified international flows of information and technology (Box 7.2) and the promotion of concepts of good governance and business management. In this con- text, efforts to develop a policy framework for shared growth beyond the construction sector, which is a major beneficiary of current invest- ment inflows, needs immediate attention. Armenia's manufacturing and services sectors should also benefit more from investments and the associated intensified international flows of information and tech- nology. Some generic recommendations are A discussion of the future agenda is contained in other sections of the documents. 178 Table 7.8 Simulation Assuming Export Diversification Part Effect on II: Effect on labor Effect on jobs household Detailed income, US$ & numbers & income, Magnitude of shock Effect on percent change percent change US$ & effected through GDP, US$ & percent Analysis policy reforms percent change Skilled Unskilled Skilled Unskilled change Total shock to economy based on the US$135 m 158.4 12.5 63.0 7,800 74,100 158.4 strategy of export diversification (7.4%) (5.5%) (7.7%) (5.4%) (7.7%) (6.4%) Vegetables and fruits, including grapes and dried fruits (9% = US$2.8 m) Crops not elsewhere classified (10% = US$2.5 m) Food processing and beverages (13% = US$2.5 m) Light manufacturing and textiles (50% = US$27 m) Metals and metal products (26% = US$22 m) Machinery equipment, motor vehicles, and precision optical equipment (34% = US$28 m) Banking, lending, and insurance (12% = US$3 m) Other services not elsewhere classified (41% = US$35 m) Source: SAM (Chapter 8 of this book). m = million. Table 7.9 An Alternate Approach That Does Not Rely on Export Growth A 10 percent positive shock in agriculture versus a 10 percent positive shock in manufacturing and industry (excluding minerals and mining) Effect on Effect on labor Effect on jobs, household income, US$ & numbers & income, Effect on percent change percent change US$ & Inward-looking policies that promote all nonmineral GDP, US$ & percent sectors equally percent change Skilled Unskilled Skilled Unskilled change Agriculture 151.3 7.2 65.2 4.6 76.6 151.3 (10% = US$91 m) (7.1%) (3.2%) (7.9%) (3.2%) (8.0%) (6.1%) Manufacturing & industry 359.9 38.6 135.8 24.5 159.7 359.9 Gr (10% = US$241 m) (16.8%) (16.9%) (16.6%) (16.9%) (16.6%) (14.6%) owth Source: SAM (Chapter 8 of this book). Analysis m = million. 179 180 Part II: Detailed Analysis Box 7.3 Information and Communications Technology Sector The ITC industry could play a leading role in Armenia's economic development. It is a versatile technology and there are no areas of prod- ucts or services where ITC cannot be applied and benefits obtained. ITC also has multiplier effects on other economic activities. Given the regional constraints to the country's trade, the significance of ITC development, especially e-commerce applications, cannot be overem- phasized. Armenia's ITC industry development program is based on its comparatively large supply of qualified ITC professionals, the relatively short durations of ITC projects, relatively low capital requirements, low wage rates, and benefits in terms of high value-added products. In the former Soviet Union, Armenia was one of the main scientific centers of the country, with special focus on the ITC industry. There were about 40 R&D companies active in the field, the biggest of which, Yerevan R&D Institute of Mathematical Machines, had about 10,000 employees and was producing both hardware (mainframes, computers mostly for Soviet military industry needs) and the corresponding software (operat- ing systems, applications). Despite the country's economic and physical hardships experienced since its independence in 1991, the ITC industry has survived and has even developed. Since 1997, the ITC industry has been advancing at an accelerated pace in all three of its subsec- tors: (i) hardware; (ii) software; and (iii) media for collection, storage, processing, transmission, and presentation of information. More than 200 companies, including about 20 American and European companies, are currently operating in the ITC sector in Armenia employing local specialists. Because job opportunities are limited, a large number of program- mers are unemployed or work as computer maintenance specialists or operators, rather than for software firms. However, software develop- ment has become so popular in Armenia that many mathematicians and physicists are changing their specialties to become software pro- fessionals. Armenia has approximately 6,000­7,000 specialists in the field, with 300 graduating every year from the Engineering University and the State University's applied mathematics faculty. The biggest software company in Armenia is HPL Armenia, a subsidiary of Silicon Valley­based HPL Inc., which employs about 150 programmers. Soft- ware exports alone amounted to an estimated US$20 million in 2000. In support of ITC development, the government declared the ITC industry as one of the priorities for Armenia's economic development by a decree adopted in December 2000. The concept of ITC industry development was endorsed by the government in April 2001, while the sector devel- opment program is currently under preparation. Source: "National Human Development Report, Armenia," UNDP, 2001. Growth Analysis 181 NOTES 1. Prepared by Ekaterine Vashakmadze, Economist, World Bank. 2. This multiplier reflects how much a unit increase in the demand for a commodity affects household income. 3. It is difficult to accurately assess the distribution of factor income among wage income, mixed income, and income from capital, as the Armenian statis- tical data on wages include only the official wage bill, which clearly underes- timates the real level of the wage bill. (See also the discussion in Box 7.1.) 4. There have been two large waves of Armenian emigration: at the begin- ning of the twentieth century and after the collapse of the Soviet Union. This reflects the present structure of the Armenian diaspora, which comprises the two types of internationally recognized migrants. The first type is recent emi- grants who are relatively young, have immediate families in Armenia, and support their families' incomes from abroad. This source of funds mostly finances consumption and is relatively stable at least over the future 20 to 30 years; from the pure accounting point of view this source is reflected as net transfers and net income from abroad. The second type of emigrant is the earlier emigrants, people whose families left about 100 years ago, and who are investing money in their country of origin if they see investment opportunities there. 5. It is estimated that at least some 20,000 people leave the country every year, mostly for economic reasons. 6. The higher the value added by labor is, the greater is the remuneration to labor and the stronger the poverty impact of an activity, if its share in all activities is large. At the same time, sectors with high employment multipli- ers have the potential to absorb a lot of employment, but the earnings from those jobs depend on the quality of jobs in the sector. The growth pattern is decisive in creating higher-wage jobs versus expanding the number of low- paid jobs. In general, low value-added agricultural activities and trade create the greatest amount of employment if one counts only the number of jobs generated per 1 million drams. However, there is a trade-off between the number of jobs and the quality of new jobs created. In contrast, the financial sector generates fewer jobs per million drams, but the quality of these jobs is superior and therefore provides sustainable sources of income and long-term security, moving more people out of poverty versus temporary employment and a reduction in transient poverty. That said, the creation of quality jobs may be constrained by the supply of skilled labor. 7. Remittances from the Armenian diaspora are a very important source of financing, signaling strong investment interest and business opportunities abroad and pointing to the potential for exploiting these opportunities back home. This source should be counted as a tremendous asset in terms of inte- gration into global markets. This source of funds primarily finances invest- ment, reacts to the macro environment and investment climate, and may be 182 Part II: Detailed Analysis highly volatile; from an accounting point of view, these are capital inflows which are debt-creating or returns-creating. 8. Unofficial remittances may be two to three times larger than official estimates. 9. The effect of the shocks on household income and GDP is identical in absolute value. Annex 7.1. Disaggregated Effect of Shocks on GDP (by sector), Armenia SAM Table 7A.1.1 Effect of a (­50%) Shock on Diamonds Base Impact Change Share in GDP (%) GDP on GDP in GDP (US$ million) (US$ million) (%) Old shares New shares Total agriculture 531.3 ­24.7 ­4.7 24.8 24.9 Total manufacturing 254.8 ­12.0 ­4.7 11.9 11.9 Total industry 191.8 ­27.7 ­14.5 8.9 8.1 Total construction 298.6 ­3.2 ­1.1 13.9 14.5 Total services 869.0 ­41.9 ­4.8 40.5 40.6 Total 2145.5 ­109.5 ­5.1 100.0 100.0 Source: Armenia SAM. Table 7A.1.2 Effect of a (­30%) Shock on Construction Base Impact Change Share in GDP (%) GDP on GDP in GDP (US$ million) (US$ million) (%) Old shares New shares Total agriculture 531.3 ­40.6 ­7.6 24.8 25.3 Total manufacturing 254.8 ­16.9 ­6.6 11.9 12.3 Total industry 191.8 ­9.8 ­5.1 8.9 9.4 Total construction 298.6 ­87.8 ­29.4 13.9 10.9 Total services 869.0 ­49.0 ­5.6 40.5 42.2 Total 2,145.5 ­204.1 ­9.5 100.0 100.0 Source: Armenia SAM. 183 184 Part II: Detailed Analysis Table 7A.1.3 Effect of a (­5.5%) Shock on Remittances Base Impact Change Share in GDP (%) GDP on GDP in GDP (US$ million) (US$ million) (%) Old shares New shares Total agriculture 531.3 ­51.5 ­9.7 24.8 23.9 Total manufacturing 254.8 ­18.6 ­7.3 11.9 11.8 Total industry 191.8 ­11.6 ­6.1 8.9 9.0 Total construction 298.6 ­2.3 ­0.8 13.9 14.8 Total services 869.0 ­55.1 ­6.3 40.5 40.6 Total 2,145.5 ­139.2 ­6.5 100.0 100.0 Source: Armenia SAM. EXPORT DIVERSIFICATION INTO MANUFACTURING AND HIGH-VALUE NONTRADABLES Table 7A.1.4 Effect of a (+US$135 million) Combined Shock Based on Relative Export Shares Base Impact Change Share in GDP (%) GDP on GDP in GDP (US$ million) (US$ million) (%) Old shares New shares Total agriculture 531.3 47.0 8.8 24.8 25.1 Total manufacturing 254.8 38.7 15.2 11.9 12.7 Total industry 191.8 10.6 5.5 8.9 8.8 Total construction 298.6 4.3 1.4 13.9 13.1 Total services 869.0 57.8 6.7 40.5 40.2 Total 2,145.5 158.4 7.4 100.0 100.0 Source: Armenia SAM. AGRICULTURE VERSUS MANUFACTURING AND INDUSTRY, EXCEPT MINERALS AND MINING Table 7A.1.5 Effect of a (+10% = US$91 million) Positive Shock in Agriculture Base Impact Change Share in GDP (%) GDP on GDP in GDP (US$ million) (US$ million) (%) Old shares New shares Total agriculture 531.3 95.6 18.0 24.8 27.3 Total manufacturing 254.8 12.0 4.7 11.9 11.6 Total industry 191.8 7.3 3.8 8.9 8.7 Total construction 298.6 2.1 0.7 13.9 13.1 Total services 869.0 34.4 4.0 40.5 39.3 Total 2,145.5 151.3 7.1 100.0 100.0 Source: Armenia SAM. Growth Analysis 185 Table 7A.1.6 Effect of a (+10% = US$241 million) Positive Shock in Manufacturing and Industry, except Minerals and Mining Base Impact Change Share in GDP (%) GDP on GDP in GDP (US$ million) (US$ million) (%) Old shares New shares Total agriculture 531.3 85.6 16.1 24.8 24.6 Total manufacturing 254.8 58.1 22.8 11.9 12.5 Total industry 191.8 31.4 16.4 8.9 8.9 Total construction 298.6 5.6 1.9 13.9 12.1 Total services 869.0 179.2 20.6 40.5 41.8 Total 2,145.5 359.9 16.8 100.0 100.0 Source: Armenia SAM. Annex 7.2 Methodology Used in the Paper T his paper depicts the Armenian subsectoral economic structure by analyzing a newly constructed 2002 Social Accounting Matrix (SAM) (see Light, Vashakmadze, and Khachatrian 2005). The results of this paper are derived from simulations based on a multiplier analysis of the SAM. The objective of our analysis was threefold: (i) to examine the vulnerabilities of the present growth pattern; (ii) to understand the channels through which growth penetrates to the labor market through employment creation and household incomes; (iii) to identify possible constraints to higher and more shared growth, which will generate sustainable jobs and lead to more effective poverty reduction. While approaching this task we have asked ourselves the follow- ing questions: (i) What are the sources of Armenia's current growth? What are their implications for jobs and household incomes? What does the structure of the Armenian economy imply in terms of sub- sector growth; (ii) How sustainable are these sources and what are the implied vulnerabilities of the current growth pattern? (iii) Are the sectors that have managed to attract investment indeed the ones with high future growth potential? (iv) What is the effect of growth on future productive assets: is the current growth pattern creating productive assets or mainly consuming foreign inflows with a lesser impact on future sustainable growth? Answering these questions will help us to understand what is needed to achieve a sustainable and balanced growth with strong spillover effects, especially on employ- ment generation and the well being of low-income households. We have applied a step-by-step approach that includes the following: 186 Growth Analysis 187 We have conducted an economic analysis that depicts a disag- gregated picture of the Armenian economy by constructing and analyzing different information sources, including the SAM, and by revealing the interrelationships between sectors as they grow. We have conducted a multiplier analysis that considers the eco- nomic impact of growth across different sectors and analyzes the linkages between growth, employment, and poverty. Based on preliminary findings, we have formed certain hypoth- eses that we have tested through a set of simulations. For this study we have developed and applied an SAM-based sup- ply-side model. The model is practical but has certain limitations; it assumes fixed prices, which is plausible in the short term for a country with highly unutilized capacity, like Armenia. The 25 sector SAM for Armenia (Light and Rutherford 2004) provides a good initial base for such an analysis. The model has been elaborated by incorporating the information on households, employment, and current account bal- ance. The 2002 SAM for Armenia is the first attempt to systematically combine data from disparate Armenian agencies in order to create a balanced, representative framework for economic analysis. It is also the first time that the Armenian survey of household consumption has been integrated together with the Armenian National Accounts. By providing a bridge between these two fundamental data sources, the SAM provides a basis for the comparison of data sources and a sound methodology for checking the consistency of the national and household accounts. Annex 7.3 Summary Notes on Labor Statistics in Armenia SOURCES OF DATA There are several NSS publications on labor statistics in Armenia: 1. Monthly reports of "Socio-Economic Situation in Armenia" are the main sources of the current situation on labor statistics, and sections of "Labor Market Indicators" generally comprise monthly data and monthly accumulative data on the following: nominal wages and salaries per worker number of economically active population by sectors number of employed persons unemployment rate 2. Data from monthly publications are used by NSS to prepare the annual Statistical Yearbook of Armenia. The yearbook is the main source of finalized and summarized labor statistics in Armenia. The sections on "Employment" and "Living Standards of Population and Social Sphere" contain the following average annual data: labor resources number of economically active population by sectors number of economically nonactive population number of employed persons number of unemployed persons registered unemployment rate unemployment benefits and average monthly data on the same nominal wages and salaries per worker 188 Growth Analysis 189 The data published in the Statistical Yearbook of Armenia are final and have priority in case of differences compared with the data on the "Socio-Economic Situation in Armenia." 3. However, the results of 2001 Census of Armenia (Armenia 2003c) may be considered a more comprehensive and detailed statisti- cal source for labor statistics in Armenia, though the census was undertaken only for a certain period of time. Along with some other valuable statistical data, the 2001 census results contain the following more useful information on "Economically Active De Jure Population" by the following categories: rural and urban occupation and employment status sex and age educational attainment type of economic activity All the indicators since 2002 represented in the section on "Employment" in the Statistical Yearbook of Armenia (excluding the officially registered unemployed) were based on data recal- culated by the results of the 2001 census. 4. Other labor surveys conducted by NSS are useful. The "Labor Force Sample Survey in the Urban Settlements in Armenia" (August­September 2001) was aimed at deriving statistical infor- mation on issues concerning sex and age, specialty composition, employment structure of the economically active population of the urban settlements in Armenia, factors influencing the latter, unem- ployment, underemployment, and the labor market in general. In this analytical publication outputs of the survey were summarized and general conclusions were drawn. Meanwhile, to have a com- plete picture, the derived indicators were compared with the offi- cial macroeconomic indicators on employment for Armenia. CONCLUSION The officially registered unemployment rates are calculated according to the Armenian law "About Employment." According to this law, the unemployed are considered persons 16 years and older who meet these conditions: do not have work have applied to the Employment State Service of Armenia to receive work and have at least one year's work experience/practice 190 Part II: Detailed Analysis Except for the officially registered unemployment rate, the labor force survey and the census are allowing the real unemployment level to be estimated. As a basis for the labor force survey and census, the International Labor Organization (ILO) recommendations on the defi- nition of unemployment were taken. According to this definition, during the reference period the unem- ployed are considered to be all persons who satisfy all three of the following criteria: do not have work or a profitable job are actively seeking work at the Employment State Service (as well as independently) are ready to start working immediately Table 7A.3.1 Selected Labor Indicator Definitions Source of definition Total labor force Employment Unemployment Statistical Economically active population Number of persons employed Officially registered unemployed are Yearbook includes all employed and includes those persons who able-bodied citizens of working age 2004 unemployed population who work at institutions and all who are resident in the territory of the develop labor markets (related organizations of all types, Republic and who in the period under to labor force demand for including those employed at review do not have work (profitable production of commodities and small organizations, public activity), seek a job with the aid services). organizations, peasant farms, of employment territorial centers, and also those engaged in and who are ready to begin work entrepreneurial activity, and self- immediately or during the period employed persons. defined by legislation. Labor Economically active population Persons were employed, did Persons were unemployed, did not Gr Force is the sum of employed and some paid work, or had some have a job during the observation owth Survey unemployed persons who form profitable employment during the week, looked for a job, and were the labor force demand in the observation week, even if it lasted ready immediately to take up work. Analysis labor market. for an hour. (continued on next page) 191 Table 7A.3.1 (continued) 192 Source of definition Total labor force Employment Unemployment Part 2001 The economically active The employed population must The unemployed comprise all those Census population (labor force) comprises satisfy one of the following two persons 15 years or older who during II: those individuals 15 years or categories: the reference period satisfy all three of Detailed older who are either employed the following criteria: 1. Paid employment or unemployed (according to a. Those who performed work a. Without work; in other words, the ILO definition) during the for wage or salary, in cash not in paid or self-employment as reference period. Analysis or kind defined in the employed population b. Those that have a job but b. Currently available for work; in were temporarily not at other words, available for paid work during the reference or self-employment, during the period, but had a formal nearest two-week period attachment to their job c. Seeking work; in other words, took specific steps in a recent period to 2. Self-employment find paid or self-employment a. Those who performed work for profit or family gain, in Also considered unemployed are those cash or kind individuals who stated they that were b. Those that have their not looking for a job due to one of the own enterprise but were following reasons: not working during the a. Have applied for a job and will be reference period for some starting it in near future specific reason b. Have applied for a job and are According to current ILO waiting for the answer recommendations, work means at least one hour of work performed during the reference period. Chapter 8 A Social Accounting Matrix for Armenia INTRODUCTION This chapter1 provides documentation related to the 2002 Armenian SAM. Development of this SAM is necessary to identify the potential linkages between economic growth and poverty alleviation and for basic economic analysis of Armenian tax and trade policies. The credibility of these studies depends upon an accurate and complete depiction of Armenia's economy. The 2002 SAM for Armenia is the first attempt to systematically combine data from disparate Armenian agencies in order to create a balanced, representative framework for economic analysis. It is also the first time the Armenian survey of household consumption has been integrated with the Armenian National Accounts. By provid- ing a bridge between these two fundamental data sources, the SAM provides a basis for comparison of data sources and a sound methodol- ogy to check the consistency of the national and household accounts. The development phase was influenced by the particular uses for the SAM. Some of these uses are itemized here: Economic analysis. A disaggregated picture of the economy is made possible by combining the national data with industry- level data. This disaggregation reveals the interrelationships among sectors as they grow. Economic growth and the incidence of poverty. By connecting poor and rich households to different factors of production, dif- ferent consumption bundles, and different transfer patterns, the household SAM provides the basis to determine what linkages exist between Armenian economic growth, employment, and the alleviation of poverty. 193 194 Part II: Detailed Analysis Multiplier analysis. The inclusion of an input-output (IO) matrix into the social accounts means that we can invert this matrix and generate economic output "multipliers." This is often a straight- forward way to consider the economic impact of growth across different sectors.2 Policy and impact analysis. This SAM provides the basis for a standardized computable general equilibrium (CGE) model of the Armenian economy. CGE models are the most common tool for tax and trade policy analysis. The underlying data describe the industries and individuals that pay taxes in Armenia and also describe who is consuming imports and who is producing exports. Data improvement. A well-documented "benchmark" dataset can be used as a starting point for further data development. We hope this SAM will be used as a benchmark to measure the quality of new data sources. Data integration. The official data produced by various Arme- nian ministries and departments, such as the NSS, Ministry of Finance, customs authorities, and the tax authorities, should be "consistency checked" with other divisions' output. Although perfect coordination is not possible, we hope this SAM will permit each organization to compare their own data with the outputs that would be applied by combining data from other sources. With these needs in mind, we developed three SAMs to represent Armenia's economy; each one reveals a different level of detail. The macro SAM represents the most aggregated national accounts. The cells from this matrix replicate the basic national accounting identities, such as GDP, national income, and the current/capital accounts. The micro SAM disaggregates total national production into 25 separate industries. Each industry has a distinct production tech- nology that includes intermediate inputs from other industries and uses different combinations of production factors. This matrix decom- poses aggregate value added into four distinct factors: skilled labor, unskilled labor, capital rents, and land rents. Total direct and indi- rect taxes are distinguished using the major Armenian tax streams: value-added taxes, income taxes, profits taxes, excise taxes, import tariffs, and payroll (social security) taxes. The international accounts are decomposed by sector, so that imports, exports, domestic supply, and total aggregate supply are known for each of the 25 industries. The micro SAM combines all of these components into a single, bal- anced matrix. A Social Accounting Matrix for Armenia 195 Finally, the household (HH) SAM projects income and expenditure profiles from a single representative agent onto 20 income deciles. Ten rural and 10 urban households are categorized by income; then each household's factor income, expenditure patterns, and transfers are mapped onto the 25 industries and four factor types. Although we believe these social accounts represent the best avail- able data for Armenia at this point, there is room for improvement. Closer collaboration with the NSS could yield a more accurate depic- tion of total sectoral output. By the same measure, closer collaboration with the NSS household survey division might yield more accurate factor incomes for each income decile. Accurate factor share informa- tion is critical in order to identify the effect of industrial growth upon wages and capital rents. Finally, an Armenian-based IO table would be an improvement because it would dispel suspicions that the core nature of Armenian production is not correctly captured. We hope the current dataset will lower the cost of further data development and lead to a universally accepted set of national accounts for the country. This documentation should facilitate further improvements, especially because a detailed description of the build- ing process has been included in the Annex. The remainder of this documentation is arranged as follows. Sec- tion 2 presents three depictions of the Armenian social accounts. Section 3 briefly describes past and future applications of the Arme- nian accounts for economic analysis. Section 4 concludes and offers potential avenues for further improvements. A lengthy and detailed Annex documents the SAM building process; Annex subsections "Matrix Balancing" and "The Build Distribution" detail the algebraic mechanics and files needed to build the SAM from scratch. THREE SOCIAL ACCOUNTS MATRICES In this section we present the Armenian accounts using three separate matrices, each with an increasing level of detail. The macro SAM con- tains only aggregate national accounts values; the micro SAM adds sectoral detail by disaggregating production, consumption, and trade to 25 goods and industries. Finally, the household SAM adds fur- ther detail by disaggregating the single consumer into 20 separate households using data from the 2002 Armenian Household Survey of income and consumption. All of the social accounts are based upon a set of input tables. These input tables are based upon published government statistics, analytical reports, and judgments from local experts and staff. 196 Part II: Detailed Analysis The Aggregate macro SAM We have chosen a specific split of accounts to reveal sufficient detail for calculating the output both from expenditure and the revenue side for constructing the macro SAM. Furthermore, we have made an effort to preserve the mandatory data entries fully comparable with the System of National Accounts. We have included the follow- ing major accounts in the macro SAM: productive sector, commodity markets, factors of production, two institutions (governance and the rest of the economy), the savings/investment account, and the "rest of the world" account. By using the column totals provided in two of the input tables, Annex Tables 8A.1.2 and 8A.1.3, we can produce most of an aggregate SAM. Several national indicator statistics from the NSS can be found in the macro SAM below. For example, GDP at producers' prices is equal to total value added, or US$2,145.5 million. The macro SAM is reproduced here as Table 8.1. Table 8.1 Aggregate Social Accounts for Armenia, 2002 (US$ millions) A B C D E F G H ACT COM VA RA GOV ROW I/S Total 1 ACT -- 4,337.6 -- -- -- -- -- 4,337.6 2 COM 1,962.1 -- -- 2,033 237 698 514.8 5,444.9 3 VA 2,145.5 -- -- -- -- -- -- 2,145.5 4 RA -- -- 2,145.5 -- 117.7 206.7 -- 2,469.9 5 GOV 230 -- -- 72 -- 54.7 -- 356.7 6 ROW -- 1,107.3 -- -- -- -- -- 1,107.3 7 I/S -- -- -- 364.9 2.0 148 -- 514.9 8 Total 4,337.6 5,444.9 2,145.5 2,469.9 356.7 1,107.4 514.8 -- Key: ACT Activities account, total production COM Commodities account, aggregate demand, and supply of goods VA Value-added account RA Representative-agent account GOV Government account ROW Foreign account. I/S Investment and savings account Source: Authors' calculations. Note: -- = not available. A Social Accounting Matrix for Armenia 197 These figures should be easy to recognize. Reading by ROW/COL- UMN, we see that total value-added (RA/VA) equals US$2,145 mil- lion. This figure is identical to the column sum for VA0 from Annex Table 8A.1.2. Total imports, found in the ROW/COM cell, are also equal to the column total under M0, US$1,107. At some point, how- ever, adjustments were impossible to avoid. For example total inter- mediate demand, ID0, is slightly larger in the macro SAM than it is in input Annex Table 8A.1.2, (US$1,962 versus US$1,924). Other cells can be checked in a similar way. Table 8.2 compares the macro SAM from Table 8.1 to officially published statistics from the NSS. There is a clear linkage between the macro SAM and the Armenian System of National Accounts. The macro SAM cells are presented in column three of Table 8.2, and the corresponding National Accounts spreadsheet cell is presented in col- umn four. Table 8.2 Replication of Selected National Accounts Measures by the Macro SAM Value SNA Account (US$ millions) Macro SAM spreadsheet cell GDP (producer prices) 2,146 A3 SNA (An7) Net indirect taxes 230 A5 SNA (An4) GDP at (market prices) 2,376 A3+A5 SNA (An3) Total absorption 2,786 D2+E2+G2 SNA (An28) GNP 2,464 A3+A5+F4+118.7 SNA (An60) GNI (disposable) 2,638 A3+A5+F4-F5 SNA (An61) GDS 105 A8-A2-D2-E2 SNA (An50) GNS 367 D7+E7 SNA (An53) GDI 515 G2 SNA (An36) Direct taxes 73 D5 GFS (AN11) Total net current transfer 173 F5+118.7 SNA (An49) Imports of goods and services 1,107 B6 SNA (An26) Exports of goods and services 698 F1 SNA (An25) Current account balance 148 F7 BOP (AN36) Source: Authors' calculations. Note: The macro SAM cell coordinates are used to calculate each macroeconomic statistic. These values can be compared with the corresponding National Accounts spreadsheet cells. SNA: System of National Accounts. 198 Part II: Detailed Analysis The micro SAM Construction of a complete SAM is done by disaggregating each cell contained in the aggregate (macro) SAM. Using more detailed tables for industrial supply (for example, see Table 8A.1.2), we can produce disaggregated statistics for production, industrial demand, value- added, final demand, and trade. The dimensions of the micro SAM for Armenia are listed in Table 8.3. The following paragraphs describe how the micro SAM accounts are organized. Activities/Commodities. We used official NSS statistics to compute participation for major sectors, but some sectors that were provided needed further disaggregation from outside data sources. For exam- ple, the NSS official statistics use "industry" to describe all manufac- turing sectors. We used output shares taken from an IMF Country Report for Armenia to further disaggregate this super-sector into six subsectors. Given total output for each sector, we then use a surrogate IO table as a starting point to describe each industry's production technology. For most sectors, we assert that the production technology is similar, agricultural industries are typically labor-intensive in low-income coun- tries, and manufacturing industries are typically capital-intensive.3 Some sectors are clearly unique to Armenia and required manual adjustment. For example, Armenia's "cut jewelry" sector accounts for one-third of total imports and exports. This sector uses only one inter- Table 8.3 Structure of the Disaggregated (micro) SAM for Armenia, 2002 Number Name of account of accounts Description Activities/commodities 25 Industries are described in Table 8.4 Factors of production 4 Skilled labor, unskilled labor, capital, and land Tax streams 6 VAT, excise, labor tax, profits tax, payroll tax, and import tariffs Government 1 Central government Households 1 One single, representative agent Trading partners 1 ROW trading partner Margins n.a. Trade and distribution (TRD) margins, collected as a production activity and used as an intermediate input Savings-investment 3 Private, government, and foreign Source: Authors' calculations. A Social Accounting Matrix for Armenia 199 mediate input, uncut gems. The remainder of the value accrues to labor and capital, so the IO matrix was corrected to reflect Armenia's special circumstances in the jewelry industry. In other cases, aggregate intermediate demand was larger than total supply. Adjustments to the IO matrix were required in these cases as well. A full description of the SAM-balancing process is included in the Annex. Tax Streams. Upon request, the Armenian Ministry of Finance has supplied figures for tax collections across tributary streams for each of the 25 commodities. Each of the six taxes is presented individually in the SAM, but where they appear in the matrix depends on whether they are classified as direct or indirect taxes. Total tax collections in our treatment are about 20 percent smaller than total government revenues for 2002. Collections are smaller in this dataset because we chose to omit some of the revenues from special taxes, fees, and other miscellaneous incomes that are not consistent year to year. Government. All government activity, local and national, is pre- sented using a single government agent, and a single government production technology. Government income comes mostly from tax collections and from transfers. Revenues are spent primarily to pur- chase the "government good," which is produced using intermedi- ate materials, labor, capital, and land. The government also transfers funds to households and contributes to savings. Households. The micro SAM includes a single representative agent (RA). The RA earns income by selling production factors, and the RA consumes goods as final demand. Total factor earnings by the RA in this matrix sums to US$2,145 million, which equals gross domestic product, taken at producer prices. Other income comes from govern- ment transfers and remittances from abroad, which are placed in the ROW account. Sector Descriptions. Table 8.4 presents 25 industrial sectors with labels and descriptions. These sectors reflect our best notion of Armenia's stra- tegic and important industries. While some of the sectors in the social accounting matrix have been disaggregated from the Armenian national statistics, others are a combination of less-important sectors that were included on separate lines in the NSS accounts. Table 8.4 presents a detailed description of each sector, and a mapping between the NSS out- put statistics and the sectors in Table 8.4 can be found in the Annex. Armenia 2002: A Snapshot The micro SAM depicts the Armenian economy at a fairly disaggre- gated level and helps us to understand the implications of its present structure in terms of subsectoral growth. Table 8.5 displays Arme- nia's production and trade statistics for 2002. The largest single out- put sectors were construction, trade, and processed foods. Other large 200 Part II: Detailed Analysis Table 8.4 Detailed Sector Descriptions for Each Industry in the 2002 Armenian micro SAM Symbol Description GRN Grains, potatoes, and legumes. Taken from the 2002 agriculture dataset (available in the Armenia SAM distribution). This sector combines wheat, potatoes, and leguminous plants. VFR Vegetables and fruits, including grapes. Includes vegetables, dried fruit (without grapes) and the grapes account. VOL Vegetable oils. Directly from 2002 agriculture accounts. OCR Other crops. Crops that are not classified elsewhere in the social accounts. MIL Milk and milk products. Dairy products including raw or processed milk, eggs, and related products OMT Other meats. Includes the beef, pork, mutton, goat, and poultry categories from the 2002 agriculture dataset. ENR Energy. Oil and natural gas. This sector is not produced in Armenia--it is a pure import. MIN Mining and quarrying. This sector comes from two sources. The first source is the National Accounts (Macro Data.xls) sector "Geology." The second portion comes from the IMF Statistical Annex, page 7, Table 4: Structure of Industrial Production (1996­2001). The portion for "Mining and Quarrying" is taken from the "industry" sector of the National Accounts. FOD Processed food and beverages. Includes all processed foods, alcoholic, and nonalcoholic beverages. This is one of the largest consumption sectors in Armenia. TBC Tobacco and tobacco products. CRP Chemicals, rubber, and plastic. This sector was apportioned from the industry macro-sector in the 2002 NSS official output statistics. Oil and natural gas refining or processing is included in this sector. OMF Other manufacturing. Manufacturing industries not elsewhere classified. LMF Light manufacturing and textiles. Includes textiles, dressing, and dying of fur. MNM Minerals and precious stones. This sector mainly captures the jewelry-cutting business in Armenia. The single largest import and export good in Armenia is jewels. Uncut jewels enter the country and finished jewelry is exported. Shares are taken from Table 4 of the IMF Statistical Annex. This datasheet is available as part of the Armenia SAM distribution (see Annex for details). MTL Metals and metal products. Basic metals and fabricated metals, taken directly from the IMF Statistical Annex. Represents 8 percent of industrial output. (continued) A Social Accounting Matrix for Armenia 201 Table 8.4 (continued) Symbol Description MCH Machinery and equipment. Combines machinery, equipment, and motor vehicles with optical, medical, and other precision devices, IMF Statistical Annex, Table 4. Combined share is 3.8 Percent. UTL Gas, water, and trash utilities. This sector appears twice: once in the industrial section from the IMF Statistical Annex, then again as a separate account in the official NSS accounts. We combine these two accounts for the total. The totals are the "utility sector" from the macro report and "electricity, gas, and water" from the IMF Statistical Annex, Table 4. ELE Electricity production and distribution. Captures nuclear electricity production and the distribution of power across the Armenian electricity grid. CON Construction. Taken directly from the official NSS National Accounts. TRN Transportation and communications. Includes public transportation activity, postal communication, and fixed-line and wireless telecommunications. TRD Trade and distribution activities. Wholesale and retail trade plus commerce combines two macro lines: Retail trade and catering plus general commerce. BNK Banking and finance. Combines two lines from the 2002 National Accounts: insurance and lending. GOV Government, defense, and public procurement. This represents government activities within the economy. Taken from the macro data spreadsheet, and also compared with general statistics regarding the government sector. This sector also includes "social spending," including four lines from the national accounts: health & sport, education, culture, and science. DWE Dwellings and housing. One line from the national accounts: housing. activities include the public sector (GOV), cut gems, transportation services, and electricity output. Food is probably more important to the Armenian economy than it appears in this table. If we combined the four largest food industries (processed foods (FOD), dairy (MIL), meats (OMT) and grains (GRN)), then consumable food would be the largest single industry in Armenia. Total combined output for these four industries accounts for 25.1 percent of total production. Table 8.5 also displays exports (X0), imports (M0), and total domes- tic supply (A0) for each sector. Import and export volume is domi- nated by a single sector: uncut and cut gems (MNM). This sector alone 202 Part II: Detailed Analysis Table 8.5 Armenian Production Statistics by Sector, 2002 (US$ million) Sector Y0 Y0(%) D0 X0 M0 A0 Construction 518.3 11.9 512.1 6.2 2.9 515.0 Retail & wholesale trade & catering 447.2 10.3 447.2 0.0 0.0 447.2 Food processing & beverages 430.7 9.9 379.6 51.1 46.7 426.3 Governance, defense, & public spending 362.5 8.4 354.8 7.7 7.8 362.6 Dairy products, eggs, & milk 342.6 7.9 342.1 0.5 6.8 348.9 Mineral products & precious stones 269.3 6.2 11.3 258.0 206.3 217.5 Transport & communication 259.9 6.0 172.2 87.7 24.0 196.2 Electricity supply & distribution 163.2 3.8 149.8 13.4 5.6 155.5 Beef, pork, mutton, & poultry meat 159.0 3.7 158.9 0.1 23.1 182.0 Wheat, potatoes, & legumes 155.5 3.6 155.2 0.3 50.7 206.0 Electricity, gas, & water supply 152.0 3.5 152.0 0.0 0.0 152.0 Vegetables, grapes, dried fruits 132.5 3.1 126.9 5.6 24.0 150.9 Metals & metal products 109.9 2.5 65.3 44.7 55.3 120.6 Crops not elsewhere classified 103.1 2.4 98.1 5.0 19.9 118.0 Services not elsewhere classified 93.7 2.2 23.1 70.6 60.5 83.6 Housing & dwellings 92.6 2.1 92.6 0.0 0.0 92.6 Machinery, vehicles, precision equipment 88.7 2.0 33.0 55.7 186.4 219.5 Tobacco products 80.9 1.9 77.3 3.6 30.0 107.3 Energy (oil & natural gas) 75.7 1.7 75.7 0.0 156.5 232.2 Mining & quarrying 65.9 1.5 27.2 38.7 1.1 28.4 Other manufacturing 61.2 1.4 56.1 5.0 52.1 108.3 Chemicals, rubbers, & plastics 53.0 1.2 44.7 8.3 81.7 126.4 Light manufacturing & textiles 52.6 1.2 23.5 29.1 40.1 63.7 Banking lending & insurance 51.6 1.2 44.9 6.7 11.3 56.2 Vegetable oils & fats 15.8 0.4 15.8 0.0 14.3 30.1 Source: Authors' calculations and NSS statistics. Key: Y0: Aggregate production value Y0 (%): Production share of total output D0: Supply to domestic market X0: Supply to export markets M0: Import value (cif) A0: Armington supply (D0 + M0) represents about one-quarter of the total trade volume for Armenia. Uncut gems are imported into the country, they are then cut and sometimes mounted as finished jewelry, then re-exported to Europe or the United States. Transportation, services, and machine products are the other large exports for Armenia. A Social Accounting Matrix for Armenia 203 Table 8.6 Value Added in Production for Armenia, 2002 (US$ million) Sector VA0 VA0(%) VA/Y L/VA K/VA Construction 298.6 13.9 57.6 40.7 59.3 Retail & wholesale trade & catering 288.6 13.5 64.5 30.2 69.8 Governance, defense, & public spending 197.7 9.2 54.5 86.1 13.9 Dairy products, eggs, & milk 168.2 7.8 49.1 55.4 44.6 Food processing & beverages 158.1 7.4 36.7 64.0 36.0 Transport & communications 153.1 7.1 58.9 57.0 43.0 Vegetables, grapes, dried fruits 115.4 5.4 87.1 48.7 51.3 Electricity, gas, & water supply 111.6 5.2 73.5 29.3 70.7 Electricity supply & distribution 102.1 4.8 62.6 48.7 51.3 Wheat, potatoes, & legumes 98.7 4.6 63.5 51.0 49.0 Crops not elsewhere classified 81.1 3.8 78.7 42.5 57.5 Housing & dwellings 68.2 3.2 73.6 0.0 100.0 Beef, pork, mutton, & poultry meat 66.9 3.1 42.0 53.0 47.0 Metals & metal products 43.6 2.0 39.7 67.9 32.1 Mineral products & precious stones 40.7 1.9 15.1 60.3 39.7 Mining & quarrying 39.4 1.8 59.8 38.6 61.4 Banking lending & insurance 35.7 1.7 69.1 68.0 32.0 Services not elsewhere classified 23.6 1.1 25.2 28.3 71.7 Other manufacturing 17.9 0.8 29.3 40.4 59.6 Tobacco products 13.7 0.6 17.0 50.4 49.6 Chemicals, rubbers, & plastics 10.4 0.5 19.6 52.2 47.8 Machinery, vehicles, precision equipment 5.8 0.3 6.5 56.8 43.2 Light manufacturing & textiles 5.3 0.2 10.1 66.2 33.8 Vegetable oils & fats 1.0 0.0 6.0 71.2 28.8 TOTAL 2145.5 100.0 0.0 0.0 0.0 Source: Authors' calculations and NSS statistics. Key: VA0: Total value added (US$ million) VA0 (%): Value added as a share of total VA0 VA/Y: Value added share of total output L/VA: Labor share of value added K/VA: Capital share of value added If we reorganize these statistics according to value added, a different story emerges. Table 8.6 contains a breakdown of factors and value added in production for each sector. The government is the most labor-intensive activity in Armenia, where labor represents 86.1 percent of total value added. Other labor- intensive industries are banking (BNK), light manufacturing (LMF), and jewelry cutting (MNM). Column three (VA/Y), displays the inten- sity of value added in production. Machinery, tobacco, and jewelry 204 Part II: Detailed Analysis cutting have the lowest share of value added in production, which implies that those industries are comprised mostly of intermediate inputs. Crops, vegetables and fruits, banking, and utilities have a rela- tively higher share of value added in the production process--growth in these sectors is more likely to contribute to higher personal income for Armenians. Some industries, such as cellular telecommunications, mining, or utilities may exhibit a high degree of value added in pro- duction (VA/Y may be large), but since these industries are capital- intensive, their contribution to personal income will depend upon who collects the capital rents. For example, a foreign-owned mine may capture significant rents, but most of this income could be for- warded offshore. Taxes Like all former CIS countries, Armenia is faced with challenges related to tax collection and public finance. Despite relatively high statutory tax rates, tax evasion and multilevel corruption are significant impedi- ments to collection. The result is a relatively weak public tax system. Table 8.7 presents Armenia's major tax streams and breaks down tax collections by sector. Value-added taxes contribute most to the public coffers: about US$158 million was collected from the VAT in 2002. The statutory VAT rate is 20 percent, and more than one-half of these revenues are collected at the Armenian border. Excise taxes on motor fuel, tobacco, and beverages collected US$56 million in 2002, labor and social secu- rity taxes accounted for approximately US$50 million, the profits tax (denoted here by TK) contributed US$22.8 million, and import tariffs amounted to approximately US$15.2 million. These main tax streams represented US$302 million in 2002.4 According to the 2004 Aide- Mémoir on Armenia, we know that 2002 tax revenues were closer to US$385 million. The collections here represent only the major tax streams, excluding several fees and additional royalties collected by the central and local governments. Consumption Household demand, government demand, and intermediate demand can help to identify how price changes, such as inflation, can have an impact on households or producers. Table 8.8 displays the total com- modity supply for Armenia (A0 = D0 + M0) and each component of demand for these goods. A Social Accounting Matrix for Armenia 205 Table 8.7 Tax Payments by Sector in Armenia, 2002 (US$ million) Tax instruments Sector Total REVS % VAT TM TL TXS TK Energy (oil & natural gas) 75.0 24.8 49.3 0.0 0.0 25.7 0.0 Food processing & beverages 32.9 10.9 16.1 2.7 3.7 9.1 1.2 Tobacco products 30.5 10.1 5.8 1.4 0.4 22.0 0.9 Machinery, vehicles, precision equipment 18.8 6.2 14.9 3.2 0.4 0.0 0.3 Other manufacturing 17.3 5.7 11.7 0.8 2.9 0.0 1.8 Retail & wholesale trade & catering 14.9 4.9 7.4 0.0 6.2 0.0 1.3 Governance, defense, & public spending 13.7 4.5 0.0 0.0 13.7 0.0 0.0 Electricity gas & water supply 12.8 4.2 4.2 0.0 2.5 0.0 6.2 Transport & communications 11.4 3.8 3.7 0.0 5.0 0.0 2.8 Chemicals, rubbers, & plastics 11.1 3.7 10.0 0.5 0.4 0.0 0.3 Wheat, potatoes, & legumes 8.7 2.9 8.7 0.0 0.0 0.0 0.0 Beef, pork, mutton, & poultry meat 7.6 2.5 3.9 1.8 0.4 0.0 1.6 Construction 7.5 2.5 2.1 0.0 3.3 0.0 2.1 Light manufacturing & textiles 6.4 2.1 4.3 1.3 0.4 0.0 0.4 Mineral products & precious stones 5.9 2.0 3.9 1.1 0.4 0.0 0.5 Vegetable oils & fats 5.1 1.7 3.5 1.6 0.0 0.0 0.0 Mining & quarrying 5.0 1.7 2.2 0.0 0.4 0.0 2.4 Housing & dwellings 5.0 1.6 0.0 0.0 4.6 0.0 0.4 Metals & metal products 4.6 1.5 3.2 0.0 1.2 0.0 0.1 Banking, lending, & insurance 3.9 1.3 0.0 0.0 3.3 0.0 0.5 Crops not elsewhere classified 2.4 0.8 1.9 0.5 0.0 0.0 0.0 Diary products, eggs, & milk 2.2 0.7 1.3 0.3 0.4 0.0 0.1 TOTAL 302.8 100.0 157.9 15.2 50.0 56.8 22.8 Source: Authors' calculations and Ministry of Finance statistics. Key: REVS: Total tax revenues for sector %: Share of aggregate tax revenues VAT: Value-added tax revenues--all sources TM: Import tariffs TL: Labor taxes TXS: Excise taxes TK: Profits taxes 206 Part II: Detailed Analysis Table 8.8 Total Commodity Supply and Component Demand by Consumer, 2002 (US$ million) Demand Sector Supply A0 C0 C0(%) G0 ID0 I0 Food processing & beverages 426.4 275.0 13.5 0.0 151.4 0.0 Dairy products, eggs, & milk 349.0 231.3 11.4 0.0 117.7 0.0 Retail & wholesale trade & catering 377.3 226.4 11.1 0.0 150.9 0.0 Energy (oil & natural gas) 232.3 131.3 6.5 0.0 101.0 0.0 Vegetables, grapes, dried fruits 150.8 109.8 5.4 0.0 41.0 0.0 Beef, pork, mutton, & poultry meat 182.1 103.3 5.1 0.0 78.8 0.0 Tobacco products 107.4 102.7 5.0 0.0 4.8 0.0 Electricity, gas, & water supply 152.0 97.5 4.8 0.0 54.5 0.0 Construction 631.6 83.9 4.1 0.0 136.3 411.4 Transport & communication 196.2 79.2 3.9 0.8 116.2 0.0 Wheat, potatoes, & legumes 205.5 75.9 3.7 0.0 129.6 0.0 Governance, defense, & public spending 316.6 73.1 3.6 230.1 13.4 0.0 Mineral products & precious stones 217.7 68.0 3.3 0.0 149.6 0.0 Electricity supply & distribution 155.4 63.0 3.1 0.0 92.5 0.0 Housing & dwellings 92.6 48.2 2.4 0.0 44.4 0.0 Banking, lending, & insurance 56.3 39.0 1.9 0.0 17.2 0.0 Other manufacturing 108.2 38.5 1.9 0.0 63.7 6.0 Crops not elsewhere classified 118.0 38.1 1.9 0.0 79.9 0.0 Machinery, vehicles, precision equipment 218.7 37.5 1.8 0.0 97.9 83.3 Light manufacturing & textiles 63.7 29.2 1.4 0.0 34.5 0.0 Metals & metal products 120.6 26.5 1.3 0.0 86.7 7.4 Chemicals, rubbers, & plastics 126.5 23.9 1.2 3.2 99.4 0.0 Vegetable oils & fats 30.1 17.9 0.9 0.0 12.2 0.0 Mining & quarrying 28.3 8.8 0.4 0.0 19.6 0.0 Services not elsewhere classified 83.5 4.9 0.2 3.0 68.9 6.8 TOTAL 4,747 2,033 100 237 0 515 Source: Authors' calculations and NSS statistics. Key: A0: Armington aggregate supply (D0 + M0) C0: Final demand by households C0(%): Budget share of household consumption by sector. G0: Government demand for commodities ID0: Intermediate-inputs (firm) demand I0: Investment demand A Social Accounting Matrix for Armenia 207 Table 8.8 is ranked according to final demand (C0). We see that processed foods, dairy products, trade margins, and fossil fuels are the largest components in final demand by households. Beyond these staples, the second-tier demand commodities are vegetables, meats, tobacco products, and basic utilities. The Household SAM The national accounts go far to explain Armenia's underlying produc- tion technology, trade position, and final demand structure. However, the use of a "representative" household agent for total demand makes it impossible to determine whether economic growth in Armenia helps to alleviate poverty. To determine the distribution of income and expenditure as the economy grows, we further disaggregate the SAM to distinguish household types. In the HH SAM, we disaggregate a single house- hold agent into 20 individual households, each distinguished by income (by deciles) and location (rural or urban). This sort of disag- gregation is made possible by using data from the annual "Survey of Living Conditions"--a survey of incomes and expenditures compiled by the NSS. Figure 8.1 shows how total sectoral expenditures are disaggregated into household-level expenditures by good. Total final demand is bro- ken down into rural and urban household spending, but the total demand still equals aggregate final demand. To do this, the 25 × 1 vector of expenditure shares, shown in the top panel of Figure 8.1, is expanded into a 25 × 20 matrix of expenditure shares by commodity and household. Likewise, aggregate value added is allocated to each household according to income decile and location. This 4 × 1 vector is disag- gregated into a 4 × 20 matrix of factor allocations. The column sums equal each household's total factor endowments and each row total equals national factor supply. The combination of row and column totals equals total value added, or GDP at producer prices. We see that this figure equals US$2,145, which is identical to value added in the macro SAM discussed early in this chapter, as well as the factor incomes from the above section on the micro SAM. The disaggregation does not come without difficulty. The most onerous problem when using household data is to reconcile house- hold consumption against household income. Several of the poorer deciles in Armenia report far lower income than they report expen- ditures. Even when government and family transfers are included, expenditure for the poorest decile in Armenia is at least 10 times 208 Part II: Detailed Analysis Figure 8.1 Expansion from a Single RA to Multiple Households expenditure shares: RA to 20 households (25×1 to 25×20) RA U1 U2 . . . R1 R10 GRN 75.9 GRN CUgrn grn grn 1 . . . . . . CR 1 CR 10 VFR 109.8 VFR . . . . . . . . . . . . . . . ... ... ... . . . . . . . . . . . . . . . DWE 48.2 DWE CUdwe dwe dwe 1 . . . . . . CR 1 CR 10 income shares: RA to 20 households (7×1 to 7×20) RA U1 U2 . . . R1 R10 LAB 817.9 LAB 2.8 8.5 . . . . . . 110.8 SKL 228.7 SKL 0.1 1.1 . . . . . . 27.8 . . . ... ... . . . . . . . . . . . . . . . GOV 117.7 GOV 5.5 5.9 . . . . . . 3.7 Source: Author. Note: Each household has a unique expenditure and income pattern, as depicted here. as large as income. As Deaton (2003) notes, this difference can be attributed to several factors. Because aggregate household income and spending must also be coordinated within the national accounts, we chose to use income shares and expenditure shares to distinguish households, then scale household totals to meet expenditures from the Living Standards Survey, and also meet aggregate final demand and factor endowments from the SAM. Table 8.9 displays income sources for each household, and Table 8.10 displays 2002 expenditures for each household. Figures from the "Total" row of Table 8.9 should be familiar by now, as they are identical to total factor incomes from the micro SAM. Total household incomes were scaled to meet total factor endowments and total (official) transfers from abroad, from the government, and invest- ment. For poor households, this leads to a significant divergence from reported income from the 2002 Household Survey. Deaton (2003) notes that this is not an anomoly only in Armenia; on page 8 he states: Income measured in the survey is on average larger than con- sumption measured in the surveys, but is in most cases less than national accounts consumption, and much less than GDP. Survey income is less than 60 percent of GDP on average." A Social Accounting Matrix for Armenia 209 Indeed, according to the Statistical Handbook (Armenia 2003b) based upon the 2002 Armenian Household Survey (Table 1, p. 51), 2002 total monthly per capita income was 12,776 drams. Of this, monetary income was 9,781 drams and nonmonetary (that is, own-produced value of consumption) monthly income was 2,995 drams. Annual monetary income, based upon the survey, is then (12 × 9,781)--117,372 drams per year, or US$204.83 per year, taken at market exchange rates. This implies that average income is US$0.56 per day. However, when Table 8.9 Household Income by Factor, 2002 (US$ million) Deciles SKL LAB CAP LND GOV ROW INV TOTAL Y/K $/DAY Rural households R1 0.0 2.5 5.1 0.6 6.0 0.0 0.0 14.2 91.3 0.3 R2 0.8 8.2 14.6 2.1 5.1 0.1 0.0 30.8 198.6 0.5 R3 2.0 12.8 21.8 3.4 5.7 0.1 0.0 45.7 294.8 0.8 R4 4.4 21.4 23.6 3.4 6.5 0.4 7.6 67.2 433.8 1.2 R5 8.4 31.3 25.4 3.7 5.6 0.9 9.6 85.0 548.5 1.5 R6 9.7 44.7 28.5 4.3 6.2 0.9 12.0 106.3 686.0 1.9 R7 12.0 52.7 39.2 6.0 4.3 2.7 14.9 131.9 850.9 2.3 R8 14.9 65.3 51.3 7.8 4.4 5.8 19.0 168.5 1087.3 3.0 R9 17.8 78.7 76.8 11.2 4.5 18.3 26.4 233.6 1507.3 4.1 R10 27.8 110.8 205.6 30.7 3.7 62.5 56.1 497.3 3208.5 8.8 Urban households U1 0.0 2.8 4.9 0.9 5.5 0.0 0.0 14.2 91.3 0.3 U2 1.1 8.5 12.3 2.9 5.9 0.1 0.0 30.8 198.6 0.5 U3 2.8 13.2 18.1 4.5 7.0 0.1 0.0 45.7 294.8 0.8 U4 5.8 19.7 21.1 4.7 7.9 0.5 11.4 71.0 458.2 1.3 U5 10.4 27.2 24.0 5.1 7.3 1.4 14.4 89.8 579.5 1.6 U6 12.3 38.0 28.6 5.9 8.1 1.4 18.0 112.3 724.7 2.0 U7 15.4 45.2 38.3 8.3 6.0 3.9 22.3 139.3 898.9 2.5 U8 19.1 56.0 49.3 10.7 6.1 8.2 28.5 178.1 1148.7 3.1 U9 23.0 69.3 69.9 15.2 6.4 23.5 39.6 246.8 1592.4 4.4 U10 41.1 109.6 167.8 41.2 5.5 76.0 84.2 525.4 3389.6 9.3 TOTAL 228.7 817.9 926.2 172.7 117.7 206.7 364.1 2834.0 Source: Authors' calculations, based upon Social Snapshot report by NSS (Armenia 2003b). Key: Figures are in millions of U.S. dollars, unless noted. R1 ... R10: Rural deciles U1 ... U10: Urban deciles SKL Skilled labor LAB Unskilled labor CAP Capital rents LND Land rents GOV Government transfers ROW Foreign remittances INV Savings demand Y/K Per-capita income $/DAY Per-capita income per day 210 Table 8.10 Final Demand Disaggregated by Household Type and Commodit, 2002 (US$ million) TOTAL R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 U1 U2 U3 U4 U5 U6 U7 U8 U9 U10 Part TRD Retail & wholesale II: trade & catering 311.4 1.9 4.7 7.0 8.2 9.3 13.2 15.2 18.4 25.6 58.4 1.7 3.8 4.0 7.8 9.1 7.9 14.0 18.6 21.9 60.5 Detailed FOD Food processing & beverages 274.9 2.2 6.1 7.9 9.3 11.1 11.6 18.1 16.4 25.3 55.6 1.1 3.2 3.5 6.5 7.0 6.7 10.5 14.5 16.1 42.1 MIL Dairy products, eggs, & milk 231.2 1.0 2.7 4.8 7.2 8.5 9.6 15.7 12.9 21.3 38.7 1.0 2.7 3.0 5.4 6.8 6.3 10.0 13.3 18.1 42.4 Analysis ENR Energy (oil & natural gas) 131.3 0.6 1.8 1.9 2.3 3.2 5.1 4.4 6.0 11.1 26.8 0.7 1.7 2.3 3.3 4.7 4.0 5.9 7.3 10.7 27.6 VFR Vegetables, grapes, dried fruits 109.8 0.5 1.4 1.9 2.5 2.7 5.8 4.4 4.4 7.0 24.4 0.5 1.5 1.8 3.2 3.8 3.4 4.9 6.9 10.4 18.5 OMT Beef, pork, mutton, & poultry meat 103.2 0.5 1.2 1.9 2.5 2.2 3.9 4.1 5.7 7.8 16.1 0.6 1.3 1.8 2.4 3.3 3.3 5.1 7.1 9.6 22.6 TBC Tobacco products 102.6 0.6 1.5 1.7 2.4 2.7 4.8 4.5 5.3 8.3 16.8 0.6 1.4 1.8 2.4 3.3 3.4 5.6 5.9 9.6 20.0 UTL Electricity, gas, & water supply 97.5 0.6 1.5 2.2 2.6 3.0 4.1 4.8 6.2 8.2 17.3 0.6 1.3 1.8 2.4 2.9 3.1 4.4 5.8 6.8 17.9 TRN Transport & communications 79.1 0.3 1.2 1.2 1.7 2.1 2.6 2.2 2.3 6.9 15.8 0.5 0.8 1.3 2.7 2.3 2.3 3.8 4.9 7.9 16.3 GRN Wheat, potatoes, & legumes 75.9 0.5 1.4 2.1 2.3 3.1 3.8 3.4 3.7 8.1 14.1 0.4 1.2 1.5 2.1 2.5 2.5 3.8 4.5 5.2 9.9 GOV Governance, defense, & public spending 73.1 1.5 1.1 1.6 1.4 1.8 0.0 3.1 5.1 7.7 22.0 0.7 0.6 2.1 0.0 0.0 1.1 5.1 3.4 14.8 0.0 MNM Mineral products & precious stones 68.0 1.5 0.0 0.0 0.0 4.3 0.0 0.0 8.7 6.7 0.0 3.3 4.8 13.5 0.0 1.3 16.2 0.9 0.0 0.0 6.9 ELE Electricity supply & distribution 62.9 0.4 1.0 1.4 1.7 2.0 2.7 3.1 4.1 5.3 11.1 0.4 0.9 1.2 1.6 1.9 2.1 2.8 3.8 4.4 11.4 Table 8.10 (continued) TOTAL R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 U1 U2 U3 U4 U5 U6 U7 U8 U9 U10 DWE Housing & dwellings 0.3 0.7 1.1 1.3 1.5 2.0 2.4 3.1 4.1 8.4 0.3 0.7 0.9 1.2 1.4 1.6 2.2 2.9 3.3 -- 8.7 BNK Banking, lending, & insurance 38.9 0.3 0.6 0.9 1.0 1.2 1.7 1.9 2.5 3.3 6.8 0.3 0.5 0.8 1.0 1.2 1.3 1.7 2.3 2.7 7.0 OMF Other manufacturing 38.5 0.2 0.5 0.6 0.8 1.0 1.3 1.8 2.6 3.7 8.6 0.2 0.5 0.6 0.9 0.9 1.1 1.9 1.7 3.0 6.6 OCR Crops not A elsewhere classified 38.1 0.1 0.1 0.6 0.8 0.9 0.2 2.9 0.7 0.0 6.5 0.3 0.6 0.7 0.8 1.4 2.1 2.3 1.9 5.4 9.8 Social MCH Machinery, vehicles, precision equipment 37.4 0.2 0.1 3.0 0.1 0.1 3.3 0.4 11.7 2.0 1.4 0.1 0.6 0.1 0.2 0.4 0.6 2.2 6.7 0.7 3.7 Accounting LMF Light manufacturing & textiles 29.2 0.1 0.4 0.5 0.8 0.6 1.1 1.1 1.7 2.8 6.4 0.2 0.5 0.5 0.7 0.8 0.9 1.3 1.6 1.7 5.6 MTL Metals & metal products 26.5 0.1 0.4 0.5 0.6 0.7 0.9 1.1 1.6 3.1 6.4 0.1 0.4 0.6 0.4 0.5 0.8 0.8 1.5 2.5 3.6 Matrix CRP Chemicals, rubbers, & plastics 23.9 0.3 1.0 1.1 0.0 0.3 0.0 1.6 0.3 3.8 4.8 0.2 0.5 0.3 0.6 2.2 1.3 0.6 0.0 5.0 0.0 VOL Vegetable oils & for fats 17.9 0.1 0.4 0.4 0.5 0.7 0.9 0.6 1.0 1.8 4.2 0.1 0.3 0.4 0.5 0.6 0.6 0.8 0.8 1.0 2.1 Armenia MIN Mining & quarrying 8.7 0.1 0.1 0.2 0.3 0.3 0.4 1.0 0.7 0.5 2.7 0.0 0.1 0.1 0.1 0.1 0.3 0.2 0.3 0.5 0.8 OSR Services not elsewhere classified 4.8 0.0 0.1 0.0 0.2 0.1 0.2 0.2 0.4 0.3 0.3 0.0 0.0 0.0 0.1 0.2 0.3 0.2 0.3 0.3 1.4 TOTAL 2,033.0 13.9 30.0 44.5 50.3 63.3 79.0 98.1 125.4 174.4 373.5 13.9 30.0 44.5 46.5 58.6 73.2 90.8 116.1 161.5 345.4 21 Source: Authors' calculations based upon Social Snapshot (Armenia 2003b) and other NSS publications. 1 Note: -- = not available. 212 Part II: Detailed Analysis using the National Accounts, we find that total value added (GDP at producer prices) is US$2.145 billion. Using 3.1 million as the national population, per capita income is US$1.90, more than three times as large as income from the Household Survey. As an initial pass, we used rural/urban shares that are approxi- mately equal. Some of the differences between rural and urban incomes can be attributed entirely to the investment account. The 2002 Household Survey includes an indicator for both rural and urban set- tlements. A straightforward improvement to this household dataset would be to replicate the urban-rural income and expenditure pat- terns more accurately. Table 8.10 presents household consumption patterns for each commodity. The household expenditure survey uses a "diary" for households to fill out, which contains a detailed list of food goods for purchase. For durable goods and nonrecurring expenses, a basic questionnaire is provided. We used a mapping procedure to connect each of the goods in the survey with the commodities in our SAM. Correctly classifying the survey goods was difficult, but a more prob- lematic issue arose when some types of goods were not covered in the survey. SAM-BASED APPLICATIONS Several studies are now possible using the multiple-household SAM. The most important research goal of this task is to identify where link- ages exist between economic growth, employment, and poverty. The next three subsections present a brief outline for future SAM-based applications. Economic Growth and Poverty Alleviation The important advantage of a SAM-based model is that it provides useful insights into the interrelationships that exist between goods and services markets, as well as the interlinkages between the demand for labor and capital in the process of production. Using the household SAM, which reflects how much a unit-increase in the demand and production of a commodity affects the demand for inputs, we can detect the vulnerabilities in the present economic growth pattern and estimate what the input requirements are to achieve sustainable and balanced growth. A detailed analysis will be presented in the separate paper on SAM-based applications. We can also calculate the economic growth patterns that affect mostly the consumption in the lowest-income deciles. As an example, consider benchmark consumption for the rural decile number four A Social Accounting Matrix for Armenia 213 (R4). Decile four is the group that is most likely to make the transi- tion away from poverty because they currently stand at or near the poverty line. Average per capita income for this decile is slightly less than US$1 per day, at US$340.70 per year. As we can see in Table 8.11 almost 60 percent of all expenditures are allocated to processed foods, dairy, meats, vegetables, and utilities. This includes 16.3 percent of expenditures which represent the retail markups paid as part of retail consumption (TRD). We can now com- pare household expenditure for R4 against the World Bank's poverty line for Armenia. For 2002, the overall poverty line, in value terms, was 12,261 drams per month, and the food-value poverty line was 7,516 drams.5 When translated to dollar values at market exchange rates, we get US$258 per year for overall poverty or US$159 per year for food-value poverty. If we add the top four food categories in con- sumption (FOD, MIL, OMT, and VFR), average annual purchases of these goods equals US$154, just US$4 less than the poverty line (see Figure 8.2). A straightforward strategy is to target policies that are most effec- tive for the lowest income deciles. These policies must either raise income or lower the real cost of consumption. For example, if the Table 8.11 Consumption and Income for Rural Household Decile No. 4 Sector R4(%) R4($) Family(4) R4(%) R4($) Family(4) FOD 18.5 60.1 240.6 SKL 8.2 28.1 112.4 TRD 16.3 52.8 211.1 LAB 40.6 138.2 552.8 MIL 14.3 46.3 185.4 CAP 44.7 152.2 608.9 UTL 5.1 16.6 66.2 LND 6.5 22.2 88.9 OMT 4.9 15.9 63.7 VFR 4.9 15.9 63.7 TBC 4.7 15.4 61.6 ENR 4.6 14.8 59.3 GRN 4.5 14.7 58.7 TRN 3.4 11.0 44.0 ELE 3.3 10.7 42.8 OTHERS 15.4 50.0 199.9 TOTAL 100 324.2 1,297 TOTAL 100 340.7 1,363 Source: Authors' calculations. Key: R4(%): Share of consumption or income as a percentage of total. R4($): Annual dollar value of income or consumption in per capita terms. Family(4): Annual dollar value of income or consumption for a family with four members. 214 Part II: Detailed Analysis Figure 8.2 Impact of Increased Wages or Lower Food Prices on Poverty for Rural Decile No. 4 10% wage increase labor 10% up income poverty line $159 $154 average food spending RD 4 $/year 77.5 155,000 poverty reduction 10% price cut income poverty line $159 $/year $153 new poverty line 155,000 77.5 poverty reduction Source: Authors' calculations. A Social Accounting Matrix for Armenia 215 return to unskilled labor (LAB) increased by 10 percent, then the aver- age individual in R4 would earn an additional US$13.82, and the aver- age household would earn an additional US$55.28. Holding all other prices constant, and assuming preferences are homothetic, the average per capita food bill would rise by US$5.91 (about 45 percent of the total increase in wages). Of the approximately 155,000 households in R4, this wage increase would pull 5,000 to 10,000 Armenian residents out of poverty. We get a similar result when the price of FOD is reduced by 10 per- cent. Given R4's baseline spending of US$60.1 on FOD, the reduction, other things being equal, is equal to US$6.01 of additional income. In both cases, the average change in income is about US$6.00 per year. All else being equal, we know that households will be better off in the first scenario, when income rises. The wage increase acts as a pure income effect, whereas the food price cut involves both an income effect and a substitution effect. Although the household can reallocate the savings under the price-cut scenario, the welfare gain is smaller because consumers will be faced with a different bundle of prices. Using this simple methodology, it would be possible to compute the feasible set of income and price changes, for all sectors and factors, which would achieve a given level of poverty reduction. This sort of analysis is not possible without a fully balanced, multiple-household SAM. Distribution of Tax Incidence Using the Armenian SAM, we can now evaluate the relative burden of taxation across household types. We can also consider how best to use Armenia's taxation and transfers system to alleviate poverty. The government collects taxes and spends the funds on public projects and on transfers to poor segments of the population. The net impact of tax reform upon households depends upon the following factors: production cost effects for firms, and the resulting impact upon employment and wages for key industries income and expenditure patterns of each household group transfer payments, as government revenues rise or fall These distributive impacts can now be determined using a CGE model. The model determines factor prices and commodity prices after a discrete shift in taxation. The resulting household income and consumption profiles can be determined as part of this equilibrium. This approach is often useful, because it captures all of the effects at the same time. For example, a government policy to increase transfers 216 Part II: Detailed Analysis to the poor may be ill advised if the additional funds are generated through a new tax on food items. The net effect may be increased transfers, but also higher food bills. The distortionary effect of new taxes then adds to the welfare burden, for all residents. Trade Liberalization and Poverty A third application of the Armenian SAM is to consider the effect of further trade liberalization in Armenia, and international trade's potential role for poverty reduction. Even though Armenia currently supports an open trade regime, where statutory tariff rates are 10 percent for most goods, the country remains highly isolated. This isolation comes from a closed border with two neighboring countries: Turkey and Azerbaijan. As the government considers the benefits and costs of a potentially open border with Tur- key, our dataset will be helpful in quantifying the net price effects for many staple goods, as well as the corresponding impact upon wages and employment for some of the more important production sectors. CONCLUSION AND DIRECTIONS FOR FUTURE IMPROVEMENT Our goal is to raise the level of policy debate in Armenia by providing a consistent, well-documented dataset that describes key production industries, household groups, international accounts, tax, and spend- ing streams. We understand that this dataset is preliminary, but we believe that policy decisions are better informed by a roughly param- eterized model, based upon the Armenian SAM, than upon the uncali- brated logic of economic theory. The dataset presents three views of the economy. The first is a top- level overview, which we call the macro SAM because it can be used to derive most of Armenia's macroeconomic indicators. The second is a mid-level view called micro SAM, which will serve as the workhorse dataset for most analysis projects. This dataset describes the interrela- tions among industries, the government, domestic households, and foreigners. Finally, we have developed a household SAM that can be used to identify the distributive effects of public policy. In many cases, a multiple-household model can be overly cumbersome for analysis, especially for tasks with a short timeframe. But there is no substi- tute for a multiple-household model to consider how government and nongovernmental policies are likely to affect the poor. In fact, for detailed poverty analysis, a much higher degree of granularity in the household accounts would be needed. We hope that future users A Social Accounting Matrix for Armenia 217 of this dataset endeavor to increase the consistency of the Household Survey data, and improve the corresponding household accounts in the Armenian SAM. We leave the reader with a short list of near-term data improvements. These improvements were chosen both because they will potentially improve the data quality for this SAM, and because they are relatively easy to carry out. Data Improvement We expect this document to be used for policy or academic pur- poses, so we invite the reader to use the Armenian SAM and to make improvements. Naturally, these improvements will only become ben- eficial if the reader intends to share and document any changes made to the dataset. Some near-term potential for data improvements are itemized below. Sectoral detail. This dataset distinguishes 25 production and con- sumption sectors. Our choice of industries had to be made at an early date, and it therefore missed some potentially important industries, either from a supply or a demand viewpoint. A useful and relatively painless improvement is to reorganize the sectoral aggregation in order to focus upon the most important sectors for a particular study. For example, if the SAM is used for a trade-policy study, identify the key import and export industries. Or, if the study is directed at poverty analysis, expand upon those goods most important for the lower-income deciles, which are mainly food, housing, and utilities. Multiple trading partners. We do not distinguish regional trad- ing partners. The single existing partner (ROW) could be disag- gregated into five or six distinctive trading regions. Household income and spending patterns. Additional analysis of household spending and income patterns will greatly improve the HH component of the Armenian SAM, without requiring any readjustment of other SAM submatrices. In the Annex we have included our mapping between goods in the Household Sur- vey and the sectors defined in the SAM. A more careful remap- ping of these goods could help create a closer fit and a clear count of Armenia's poor population. A separate improvement is to account for Armenia's very rich population. We believe that this tiny segment of the population (perhaps 1 percent), could account for more than 20 percent of the country's spending and more than 50 percent of the country's income. Without an assess- ment of this population segment, we will always find a yawning gap between household income and GDP. 218 Part II: Detailed Analysis Production statistics. Although output for some sectors is care- fully measured (electricity production is an example), other sectors report output that clearly conflicts with other national accounts, such as imports and exports. Additional clarification of what the NSS reports, and whether pertinent information is missing, would help to identify those industries that are truly important to Armenia's economic growth. Annex 8.1 Social Accounts Construction--A Manual T his Annex contains a detailed description of the SAM develop- ment process. The process begins with construction of several input data tables of industrial production, consumption, trade, and taxation. These input tables are combined and compared against some standard accounting identities. Where large discrepancies arise, basic judgment is combined with matrix-balancing techniques to impose a balanced set of accounts that resembles Armenia's economy. The resulting industrial data are then merged with data from the House- hold Survey into a single, balanced SAM for industries as well as households. First we discuss some challenges we encountered; then we elaborate on the actual process of building the SAM in subsequent sections of the Annex. SAM BALANCING AND MISSING DATA Development of a balanced SAM typically requires adjustments that reconcile disparities within existing data, but Armenia is the first case where at least one part is wholly missing. To complete the social accounts, we used a surrogate input-output (IO) table from another country (Poland). Of course, the new IO table did not fit perfectly with aggregate output values for Armenia. We found, however, that the production technology for most sectors is similar across develop- ing countries. A few sectors, such as jewel cutting for Armenia, are unique and require special consideration. When balancing a SAM, we attempt to utilize as much information as possible and adjust the data by weighting credible and accurate sources most heavily. Data elements that are less credible are adjusted more intensively than 219 220 Part II: Detailed Analysis others. Two techniques for SAM-balancing seem to be popular: RAS and maximum entropy (ME).6 We believe that the choice of balancing minimand (such as ME or least squares) is not as important as the choice of which data elements are being adjusted and why. Our experience has been that the study results are more sensitive to the policy variables, rather than IO coefficients. For example, results from a tax policy study are more biased by errors in the tax rate than by an equal percentage error in the IO coefficients. Therefore, when we balanced the 2002 Armenian SAM, the tax rates were held fixed, while the IO coefficients and other data were adjusted according to traditional Bayesian priors. A Structured Build Approach In the process of developing the 2002 Armenian social accounts, we tried to design a "structured build process" for developing countries where some data are missing. We hope that the process developed here can be refined and applied to new countries rapidly and cost- effectively. The approach requires access to the Global Trade Analysis Project (GTAP) database,7 competence with the General Algebraic Modeling System (GAMS) programming system,8 and with Microsoft Excel. The process does not eliminate the need for local and international exper- tise and judgment, but it can help to speed up the difficult task of SAM deployment. The approach first requires basic data collection from the host country, to be presented as a set of input tables. These tables are then combined, and if necessary, augmented with surro- gate data to generate the SAM for the target economy. This process is described in detail earlier in this chapter. DATA COMPILATION Only a limited number of data tables are required to generate SAMs. We now present these input tables and describe how they are com- bined and adjusted as the dataset is built. We emphasize that these input tables should be straightforward to construct by local experts, and that they are probably best left to these experts because they require an intimate understanding of the local economy and local accounting practices. Data acquisition and organization can be the most time-intensive portion of the process. In the Armenian case, it took several months to construct the input tables, and only a few weeks to combine the tables into a consistent SAM. A Social Accounting Matrix for Armenia 221 INPUT DATA The required input data are listed below. The data were collected, reviewed several times, and then placed into Excel (XLS) worksheets. We present an enumerated list of the required inputs: 1. Sectoral output (Y0), value-added (VA0), and intermediate demand (ID0). This is basic data that should be supplied by the country for each sector or commodity in the SAM. This input table was provided as a spreadsheet. It has been replicated here as Table 8A.1.2. 2. Sectoral imports (M0) and exports (X0). International trade sta- tistics are almost always available at the highest level of dissa- gregation, and the trade data will usually need to be aggregated to match the sectoral structure in the SAM. Aggregating trade values is simple, but aggregating the corresponding tariff rates is more difficult. For this SAM, we used a trade-weighted average of import tariffs. The tariff was weighted by the relative share of each sub-sector during the aggregation. Bach and Martin (2001) show that the trade-weighted average tariffs will under-estimate the true tariff-equivalent in a CGE model. They do not, how- ever, develop a useful alternative measure for the applied CGE modeler. Thus, the trade values are included together with other sectoral data into Input Table #1. The tariff values are included separately, in Table 8A.1.3. 3. Collected taxes. Collections, by sector of each type of tax in the economy. At a minimum, this would include import duties, labor taxes, payroll taxes, value-added taxes, and excise taxes. Vari- ous intermediate taxes, production levies, and corporate taxes may be available. Most importantly, these taxes should be sup- plied by sector as well as tax stream. This is also provided in spreadsheet format, but has been included in this report as Table 8A.1.3. 4. Intermediate input coefficients. Armenia does not have an IO table, so the IO coefficients were constructed using representa- tive values taken from the GTAP database. Poland was used for the surrogate data. Despite the Armenian conventional wisdom, we have found in a previous study that tax-policy results are not highly sensitive to IO coefficients, and that another (similar) country's production technology could be used. 5. Value-added shares. Like the IO coefficients, these values were only partially available for Armenia. We utilized local data where possible, then filled in the gaps using the GTAP database. 222 Part II: Detailed Analysis 6. Consumption shares. Government expenditures are almost always available, but they are listed by functionary stream, not by production sector. Investment demand by firms, households, and the government also lacked specific sectoral detail. House- hold demand can be computed as the residual demand after government, investment, and exports, but it is useful to consider final demand shares from different economies as a comparison. 7. Macroeconomic and financial totals. Net public and private investment, borrowing, net trade positions, international remit- tances, and other financial transactions are not crucial when con- structing a CGE model, but accountants and politicians tend to focus upon these figures. Therefore, it is best to include them in the aggregate SAM for completeness. In our case, the financial transactions and institutional transfers are found in the Macro SAM presented as Table 8.1. This Annex includes five input tables for the SAM, plus two addi- tional tables required to construct the multiple-household SAM. Input tables #1 to #4 are presented in this section of the Annex. Input Table #5, the macro SAM, is listed at the beginning of the report, as Input Table 8.1. Input tables #6 and #7 are also printed in the main body of the chapter. They are Tables 8.9 and 8.10, respectively. These tables are also included in the distribution archive as Excel spreadsheets. The data-flow process is depicted using a flow chart in Figure 8A.1. This chart shows how each of the data tables is used to generate either the micro or HH SAM. ECONOMIC VARIABLES AND THEIR CONSTRUCTION Since the SAM will be used as the basis for economic research, we may as well begin considering the data using economic terminology. Identifying these data in economic terms will help to streamline the data-automation process later. During the build process, these variables are assigned to each account in the input data, then combined to build the SAM. The construction process in which these variables are used is described next. ACCOUNTING IDENTITIES In order for the overall SAM to be balanced, we must impose a basic set of accounting identities. These identities must hold across all of the input tables before construction of the SAM. It is the responsibil- ity of the local expert to ensure that the input tables conform to the accounting identities presented in this section. A Social Accounting Matrix for Armenia 223 Table 8A.1.1 Input Tables Required for the Structured Approach to SAM Development Table Description Table #1 Sectoral production and trade values for Armenia (Table 2A.1.2) Table #2 Collected taxes by sector and tax stream (Table 2A.1.3) Table #3 Surrogate IO table for Armenia (Poland) (Table 2A.1.4) Table #4 Sectoral value-added and consumption shares (Table 2A.1.5) Table #5 Aggregate social accounts for Armenia (Table 2.1) Additional household tables Table #6 Expenditures for each household and each commodity (Table 8.10) Table #7 Incomes for each household, by factor and transfer type (Table 8.9) Figure 8A.1 Simplified Structure of SAM Development Process Macro-SAM Output Taxes IO Table recalib.gms sam.gms Micro-SAM HH-CONS HH-expend ReadHH.gms HHsam.gms HH-SAM Source: Author. Note: Three social accounting matrices are generated in this process: macro-SAM, micro-SAM, and HH-SAM. The input data enters at one of two points in the process, either before or after the micro-SAM is built. Four main GAMS (.gms) programs are used to combine the data in each table and accommodate the various discrepancies between each data source. 224 Part II: Detailed Analysis Table 8A.1.2 Input Table #1: Sectoral Production and Trade Values for Armenia, 2002 (US$ million) Sector Y0 VA0 ID0 M0 X0 D0 GRN 156 103 53 51 0 156 VFR 144 126 18 24 6 138 VOL 1 1 1 18 0 1 OCR 102 84 18 20 5 97 MIL 346 168 178 7 1 346 OMT 155 67 88 23 0 155 MIN 58 39 19 1 42 16 ENR 0 0 0 165 0 0 FOD 406 158 248 47 51 355 TBC 48 14 34 30 4 44 LMF 42 5 37 42 30 13 MNM 268 41 228 214 259 9 CRP 35 10 25 85 8 27 MTL 104 43 61 56 45 59 MCH 64 6 58 164 56 7 OMF 42 18 25 49 5 37 ELE 170 108 62 6 13 157 UTL 146 109 37 0 0 146 CON 516 298 218 3 6 510 TRN 247 144 103 24 82 165 TRD 432 266 166 0 0 432 DWE 91 66 25 0 0 91 BNK 49 36 13 11 7 42 OSR 85 24 61 61 71 14 GOV 362 211 150 8 8 354 TOTAL 4,070 2,145 1,924 1,107 699 3,371 Source: Author's calculations based upon NSS statistics. Y0: Total production for Armenian sectors VA0: Total value-added in production by sector ID0: Total intermediate demand by sector M0: CIF import value by sector X0: FOB export value by sector D0: Armenian production that is sold to the domestic market: by definition, D0 = Y0 ­ X0 Descriptions for each sector listed in this table can be found in Table 8.4. A Social Accounting Matrix for Armenia 225 Table 8A.1.3 Input Table #2: 2002 Collected Taxes by Sector and Tax Stream VAT collections (US$ million) Other tax collections (US$ million) Sector VATP VATS VATM VATD TXS TM TL TK TSS ENR 0 0 22.11 31.75 28.07 0 0 0 0 81.9 OMF 6.84 0 2.11 3.86 0 0.88 1.23 2.46 2.67 17.4 FOD 0 0 8.42 9.12 1 2.98 1.58 1.58 3.44 33.7 TBC 3.33 0 2.81 0.18 24.04 1.58 0.18 1.23 0.38 33.3 MCH 0 0 14.74 1.58 0 3.51 0.18 0.35 0.38 20.4 TRD 3.33 2.81 0 1.93 0 0 2.63 1.75 5.73 12.5 CRP 0 0 9.30 1.58 0 0.53 0.18 0.35 0.38 11.9 TRN 1.58 0.53 0 1.93 0 0 2.11 3.68 4.59 9.8 GRN 0 0 9.47 0 0 0 0 0 0 9.5 OMT 0 0 4.21 0 0 1.93 0.18 2.11 0.38 8.4 UTL 0.70 0 0 3.86 0 1.05 8.25 0 2.29 13.9 LMF 0 0 3.86 0.88 0 1.40 0.18 0.53 0.38 6.8 CON 0 0 0 2.28 0 0 1.40 2.81 3.06 6.5 MNM 0 0 2.81 1.40 0 1.23 0.18 0.70 0.38 6.3 MIN 0 0 0.18 2.28 0 0 0.18 3.16 0.38 5.8 GOV 0 0 0 0 0 0 5.79 0 12.61 5.8 VOL 0 0 3.86 0 0 1.75 0 0 0 5.6 MTL 0 0 1.58 1.93 0 0 0.53 0.18 1.15 4.2 OCR 0 0 2.11 0 0 0.53 0 0 0 2.6 DWE 0 0 0 0 0 0 1.93 0.53 4.20 2.5 MIL 0 0 1.40 0 0 0.35 0.18 0.18 0.38 2.1 BNK 0 0 0 0 0 0 1.40 0.70 3.06 2.1 VFR 0 0 1.05 0 0 0.35 0 0 0 1.4 SOC 0 0.53 0 0 0 0 0.88 0 1.91 1.4 TOTAL 16 4 90 65 62 17 22 31 48 353.6 Source: Authors' calculations based upon Ministry of Finance statistics. Definitions: Presumptive (VATP): VAT revenues collected via the presumptive tax legislation Simplified (VATS): VAT revenues collected via the simplified tax legislation Customs (VATM): VAT revenues collected at the Armenian border by customs authorities Domestic (VATD): VAT revenues collected via the traditional VAT system Tariffs (TM): Tariff revenues collected at the Armenian border Income (TL): Income tax revenues from individuals filing and from paycheck withholdings Profits (TK): Corporate profits tax revenues Payroll (TSS): Social security and other earmarked taxes applied to labor income Table 8A.1.4 Input Table #3: Production Technology Structure for Armenia (%) 226 Sector OCR OMT VOL MIL CRP ELE OMF TRD DWE GRN VFR ENR MIN FOD TBC LMF MNM MTL MCH UTL CON TRN BNK GOV OSR OCR 30.4 19.8 2.9 17.6 0.1 -- 7.2 0.7 -- 0.9 3.2 -- -- 8.9 14.9 0.4 -- 0.1 0.0 -- 0.0 -- -- 0.3 0.0 Part OMT 6.1 1.8 18.5 2.0 -- -- -- 0.7 -- 10.5 3.3 -- 0.3 19.7 0.7 0.0 -- -- -- -- -- -- -- 0.4 -- VOL -- 7.0 50.0 4.4 0.1 -- -- 0.4 -- -- 0.4 -- -- 2.5 0.6 0.0 -- -- -- -- -- -- -- 0.2 0.0 II: MIL 3.0 7.7 0.6 29.4 0.1 -- -- 2.7 -- 5.7 1.6 -- -- 0.3 0.3 0.1 -- -- 0.0 -- -- -- -- 3.4 0.2 Detailed CRP 6.0 4.6 7.4 8.0 45.6 0.6 6.8 8.1 -- 27.5 13.3 9.6 1.5 4.3 3.5 12.4 14.6 9.8 9.8 1.5 6.5 5.2 0.5 4.5 3.5 ELE 1.3 2.2 0.8 2.1 4.6 37.1 2.5 0.9 -- 2.5 3.3 -- 10.8 1.0 0.7 1.8 5.9 6.9 1.4 5.7 0.3 0.9 1.5 -- 5.4 OMF 14.1 1.6 2.4 4.7 6.9 2.4 41.0 6.2 -- 8.3 23.0 4.4 1.6 2.9 5.6 1.9 5.4 4.9 1.6 0.9 6.0 7.2 6.4 4.7 8.0 TRD 4.6 2.4 1.2 2.3 4.0 1.1 3.0 29.1 -- 4.2 5.3 5.8 2.0 2.0 2.7 2.0 4.4 2.1 5.5 2.4 7.3 13.3 24.3 16.7 10.4 Analysis DWE -- -- -- -- -- -- -- 0.0 -- -- -- -- -- -- -- -- -- -- -- -- 0.0 0.0 -- 0.0 0.0 GRN 0.0 29.8 0.1 6.1 -- -- -- 0.7 -- 15.5 0.0 -- -- 6.3 8.0 0.0 -- -- -- -- -- -- -- 0.5 -- VFR -- 0.1 0.5 0.2 0.1 -- -- 2.2 -- -- 7.7 -- -- 7.4 0.8 0.0 -- -- -- -- -- -- -- 0.9 -- ENR 2.0 0.2 -- 0.0 3.6 51.7 18.1 -- -- -- 9.8 -- 1.6 0.4 0.3 0.5 6.5 3.2 0.2 15.0 0.2 -- -- 0.4 0.7 MIN 0.4 -- 0.3 0.1 0.5 -- 0.1 0.1 -- 0.6 1.5 1.1 68.3 0.0 -- 0.1 2.2 15.0 0.0 -- 1.3 0.1 -- 0.3 0.6 FOD 0.6 12.0 5.6 5.0 0.4 -- 0.1 7.1 -- -- -- -- -- 34.4 23.6 1.5 -- -- -- -- -- 0.1 0.3 3.5 0.7 TBC 0.1 -- 0.3 0.1 0.1 -- -- 0.8 -- -- -- -- -- 0.3 20.0 -- -- -- -- -- -- 0.1 0.4 0.4 0.7 LMF 1.4 0.6 0.3 1.9 4.6 -- 3.3 4.2 -- 0.1 2.3 3.1 -- 0.5 1.6 64.5 2.4 0.5 0.9 0.1 1.0 1.4 0.1 4.9 1.4 MNM 3.8 0.2 0.0 0.4 1.0 -- 1.0 0.9 -- 1.2 3.4 1.1 0.8 0.5 1.9 0.4 26.0 2.4 0.8 1.0 20.0 0.6 0.1 1.3 3.5 MTL 6.5 1.1 0.3 2.2 1.8 0.2 4.0 5.9 -- 1.2 4.3 15.7 1.2 0.5 1.1 0.5 6.2 40.3 16.8 6.3 11.8 2.0 -- 1.0 2.0 MCH 9.4 3.4 5.0 4.9 2.1 2.3 2.4 7.1 -- 10.2 2.8 34.5 4.2 2.0 2.2 1.7 8.1 5.8 53.2 2.0 10.5 18.2 2.4 6.5 6.9 UTL 0.3 0.1 0.0 0.1 0.2 0.1 0.1 0.2 -- 0.2 1.5 0.4 0.1 0.2 0.1 0.2 0.2 0.2 0.2 57.1 0.1 0.1 0.3 0.7 1.1 CON 1.3 0.8 0.1 0.7 2.6 0.9 ­0.5 1.5 100 2.4 1.0 7.3 0.7 0.3 0.6 0.4 1.7 1.2 0.7 2.8 21.6 1.7 -- 20.2 5.4 TRN 4.8 1.7 1.3 2.4 5.8 1.7 4.3 8.6 -- 3.1 4.0 4.8 5.6 2.8 2.7 2.2 8.1 2.9 2.7 1.7 5.2 38.7 11.0 8.1 6.1 BNK 0.8 0.4 0.8 1.3 1.1 0.4 1.9 1.2 -- 1.0 1.4 -- 0.2 1.2 2.9 3.5 2.4 1.2 1.6 0.4 0.6 1.1 34.5 1.8 0.6 GOV 0.4 0.6 -- 0.7 0.7 0.1 0.2 1.1 -- 0.2 0.7 1.3 -- 0.2 0.4 0.5 0.3 0.1 0.4 1.0 0.3 0.6 3.4 6.7 1.3 OSR 2.7 1.7 1.4 3.1 13.7 1.5 3.4 9.5 -- 4.6 6.2 10.9 1.1 1.6 4.7 5.4 5.8 3.5 3.9 2.1 7.2 8.7 14.8 12.7 41.5 TOTAL 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 Source: Authors' calculations based upon GTAP dataset. Note: This table displays the production technology for a surrogate country (Poland) that constitutes the starting point for Armenia's assumed production technology. The final IO shares differ substantially for some sectors (such as the cut-jewels sector) based upon local knowledge. Other changes are required in order to satisfy the basic accounting identities listed below. -- Not available. A Social Accounting Matrix for Armenia 227 Table 8A.1.5 Input Table #4: Value Added and Consumption Shares (%) Sector LAB SKL CAP LND Total C G I Total OCR 43.2 1.5 28.1 27.1 100 100 0.0 0.0 100 OMT 52.3 0.7 11.9 35.1 100 100 0.0 0.0 100 VOL 61.6 10.4 19.4 8.6 100 100 0.0 0.0 100 MIL 52.8 2.6 21.3 23.3 100 100 0.0 0.0 100 CRP 41.3 10.6 48.1 0.0 100 80.9 19.1 0.0 100 ELE 32.4 13.6 54.0 0.0 100 100 0.0 0.0 100 OMF 34.8 5.2 60.0 0.0 100 82.5 0.0 17.5 100 TRD 20.8 3.4 75.8 0.0 100 100 0.0 0.0 100 DWE 0.0 0.0 100 0.0 100 0.0 0.0 0.0 0.0 CGD 0.0 0.0 100 0.0 100 0.0 0.0 0.0 0.0 GRN 52.2 0.8 12.0 35.0 100 100 0.0 0.0 100 VFR 52.2 0.8 12.0 35.0 100 100 0.0 0.0 100 ENR 51.3 3.7 4.4 40.6 100 100 0.0 0.0 100 MIN 33.6 4.9 47.9 13.5 100 0.0 0.0 0.0 0.0 FOD 54.0 9.9 35.1 1.0 100 100 0.0 0.0 100 TBC 44.0 6.4 49.6 0.0 100 100 0.0 0.0 100 LMF 57.8 8.4 33.8 0.0 100 100 0.0 0.0 100 MNM 52.0 8.3 39.7 0.0 100 100 0.0 0.0 100 MTL 58.4 9.5 32.1 0.0 100 28.8 0.0 71.2 100 MCH 45.0 11.9 43.2 0.0 100 31.8 0.0 68.2 100 UTL 21.3 8.7 70.0 0.0 100 100 0.0 0.0 100 CON 35.4 5.5 59.2 0.0 100 14.5 0.0 85.5 100 TRN 44.4 9.9 45.7 0.0 100 98.3 1.7 0.0 100 BNK 36.1 32.0 32.0 0.0 100 100 0.0 0.0 100 GOV 33.6 53.4 13.0 0.0 100 15.8 84.2 0.0 100 OSR 15.0 13.3 71.7 0.0 100 73.8 11.5 14.7 100 Source: Authors' calculations from NSS and GTAP database. Notes: LAB: Unskilled labor SKL: Skilled labor CAP: Capital LND: Land rents C: Final demand by households G: Government demand I: Investment demand 228 Part Table 8A.1.6 Symbol Table II: Symbol Description File location Detailed Yi Total output supply. Source: NSS National Accounts, subtables from the IMF Statistical Annex Armenia.xls (2002), and the Agriculture subtable. Xi, Mi Imports and exports of goods and services. Provided for 2002 by the customs authorities and Armenia.xls Analysis the NSS. Each traded good was mapped onto the model sectors. Di Domestic supply. This is an imputed parameter that depends upon the values for Yi and Xi. n.a. Ai Armington aggregate supply. Ai represents the total supply for a good in Armenia. It is a n.a. computed parameter, which combines Di and Mi, net of import and value-added levies applied at the border and domestically. Although this value is not incorporated directly into the SAM, it is a common input for CGE models. IDji Intermediate demand. Aggregate values for intermediate demand are provided in the National Armenia.xls; Accounts data. But individual input coefficients are taken from a surrogate IO table. This IO GTAP database table is reproduced for the reader in Table 2A.4. Li, Ki Labor and capital demand for production in sector i. Total value added is provided in the Armenia.xls; National Accounts, but the share of labor, capital, land, and indirect taxes is not provided by GTAP database the NSS. These values are based upon the surrogate IO tables from the GTAP database. The capital intensity can be an important determinant of labor/capital returns and should be reviewed more carefully by experts in Armenia. Table 8A.1.6 (continued) Symbol Description File location fi Factor endowments. Factor endowments for labor, capital, and land are computed from total n.a. demand by firms and government. Total sales of labor and capital are then allocated to households based upon the 2002 household consumption and income survey. A data discrepancy exists between official wage statistics and the official value added statistics. Although total value added is reported to be more than US$2,100 million, officially reported wages are only (approximately) US$600 million, A which implies that labor's share in value added is less than 30 percent of total value added. The Social typical labor share in value added is closer to 70 percent. Gi Government demand for good i. These values are based upon the GTAP surrogate dataset, GTAP database Accounting but can be reconciled also by reports by the Armenian Ministry of Finance. INVi Investment demand for good i. Taken from the surrogate dataset from GTAP. World Bank LDB; Armenia.xls FDi Final demand for good i. Computed as a residual based upon total supply and total demand. World Bank LDB; Matrix Total final demand should be compared with the surrogate data, as well as with Armenian Armenia.xls household surveys. This task is forthcoming in a future study. ti Direct and indirect taxes. Tax rates are computed as the ratio of collections to tax base. For example, Armenia.xls for applied import tariffs for good i equal duties for good i, divided by imports for good i: ti M = TM0i Armenia . M0i Note: Each symbol is listed together with a brief description and file location. n.a. Not applicable. 229 230 Part II: Detailed Analysis The accounting behind any SAM reflects the basic economics of pro- ducer competition, market equilibrium, and resource exhaustion. For pro- ducers, the output (sales) value must be at least as large as the combined input cost; and when the sales value exceeds input costs, this excess is often attributed to undocumented or unreported costs, such as the return to capital or depreciation. The supply of goods must be sufficient to sat- isfy total demand. Total supply for a commodity is reported as the col- umn total, and demand for a commodity is reported in the correspond- ing row total. A difference in these totals indicates a market imbalance. Finally, households in the economy earn money either by selling factors or through government and foreign transfers. The government collects revenue by levying taxes. Each agent's income must be sufficient to cover the purchases for that year. If there is an imbalance, this value is attrib- uted to either savings or borrowing. In all three situations, the data are considered to have been the result of optimizing economic agents. These basic conditions are understood more clearly if we present a handful of accounting identities related to individual SAM accounts. First, we take the set I to represent individual production sectors in the model and the corresponding output commodity.9 For example, I is the set {GRN, MIN, FOD, etc...}. We define the following conditions for producer profits and market equilibrium:10 The total sales value must equal the total cost of production: Yi = IDji + Li + Ki + Ti (8A.1.1) j Total output or sales for good i (Yi), at producer prices, must be large enough to cover the cost of production. This includes the purchase of intermediate inputs (IDji), value added (Li, Ki), and taxes (Ti). Total supply must equal demand for all commodity markets: Yi + Mi IDij + Gi + FDi + INVi + Xi (8A.1.2) j where total supply in this framework equals total output (Yi) plus imports (Mi) and demand comprise intermediate demand by firms (IDij), government demand (Gi), final consumer demand (FDi), invest- ment demand (INVi), and demand by the rest of the world (Xi). The same condition holds for factor markets. Supply of labor and capital must be sufficient to satisfy producer demand: Lh Di L (8A.1.3) h i Kh Di K (8A.1.4) h i where Lh is each household (h)'s endowment (or supply) of labor, and Kh is each household's capital endowment, and Di is the demand for L labor by sector i. So, total factor supply equals factor demand. A Social Accounting Matrix for Armenia 231 DATA MAPPING An inevitable, and unenviable, task is to map various sector types between each other. A mapping that aggregates many sectors onto a single sector is fairly easy. Values are added together and relabeled. A mapping that disaggregates a single sector into many smaller sectors is much more difficult. Additional information must be uncovered from a separate source in order to generate a useful disaggregation. This disaggregation problem was encountered with the Armenia data when mapping the industry megasector into several smaller indus- trial units. We used output shares within production to identify each subsector's production, but the relative share of intermediate inputs, labor inputs, and capital inputs were not available, so each subsec- tor uses an equivalent split between intermediate inputs and value added. Of course, we know these are different. The sector mappings between the NSS National Accounts, GTAP, and the 2002 Armenian Household Survey are described in this section. Soviet Accounts to National Accounts As of 2004, the NSS has been preparing to develop a system of national accounts based on the United Nations' International Stand- ard Industrial Classification (ISIC). Unfortunately, the 2002 statistics still rely upon the old Soviet-style system of accounts. In the old-style accounts, industry is listed as a single activity, which comprises one- third of total GDP. At the same time, small and meaningless sectors are included, like geology, which represents less than 1 percent of GDP, and has no clear corresponding economic sector. Table 8A.1.8 presents a mapping that we chose between the NSS GDP accounts and the sectors which are believed to be representative of Armenia's economy. Where possible, the percentage breakdown is shown next to the sector abbreviation. We aggregate and disaggre- gate the data in both directions. Industrial activity is disaggregated into several smaller activities, while health, education, and science are combined into a single sector, GOV. This mapping is available elec- tronically as part of the Armenian SAM distribution. The spreadsheet is named mapping.xls. GTAP Database to Armenian Industries In order to use the GTAP surrogate IO data together with the Armenian NSS data and outside data, we are required to map GTAP codes onto each 2002 Armenian SAM sector. Table 8A.1.7 presents this mapping. The left-hand column contains the three-letter Armenian code, the center 232 Part II: Detailed Analysis column is the GTAP Version 5 code, and the right-hand column is a short description of the GTAP sector being aggregated. The interested reader can further identify GTAP's sectoral classification, and a mapping between GTAP and the ISIC (International Standard Industrial Classifi- cation) accounts, on the GTAP Web site (http://www.gtap.org/). Table 8A.1.7 Mapping GTAP V5 Sectors onto 2002 Armenian SAM Sectors Armenian SAM GTAP GTAP description GRN PDR Paddy rice WHT Wheat GRO Cereal grains nec VFR V_F Vegetables, fruit, nuts VOL OSD Oil seeds VOL Vegetable oils and fats OCR CB Sugar cane, sugar beet PFB Plant-based fibers OCR Crops nec FRS Forestry SGR Sugar OMT CTL Bo horses OAP Animal products nec MIL RMK Raw milk MIL Dairy products FOD FSH Fishing CMT Bo meat products OMT Meat products PCR Processed rice OFD Food products nec TBC B_T Beverages and tobacco products ENR COL Coal OIL Oil GAS Gas MIN OMN Minerals nec LMF WOL Wool, silk-worm cocoons TEX Textiles WAP Wearing apparel LEA Leather products OMF LUM Wood products PPP Paper products, publishing P_C Petroleum, coal products OMF Manufactures nec CRP CRP Chemical, rubber, plastic products MNM NMM Mineral products nec MTL I_S Ferrous metals NFM Metals nec FMP Metal products A Social Accounting Matrix for Armenia 233 Table 8A.1.7 (continued) Armenian SAM GTAP GTAP description MCH MVH Motor vehicles and parts OTN Transport equipment nec ELE Electronic equipment OME Machinery and equipment nec ELE ELY Electricity UTL GDT Gas manufacture, distribution WTR Water CON CNS Construction TRD TRD Trade TRN OTP Transport nec WTP Water transport ATP Air transport CMN Communications BNK OFI Financial services nec ISR Insurance OSR OBS Business services nec ROS Recreational and other services GOV OSG Public administration, defense, education, health DWE DWE Ownership of dwellings Source: Armenia SAM. SURVEY CONSUMPTION CATEGORIES TO NATIONAL ACCOUNTS INDUSTRIES AND FACTORS Similar to the industrial and income classifications, we mapped house- hold consumption categories from the 2002 Armenian Household Sur- vey onto the 2002 Armenian SAM sectors. There are 480 consumption categories distinguished in the Household Survey--most of the food items. Table 8A.1.9 presents an excerpt of this mapping. MATRIX BALANCING In order to finish with a balanced set of national accounts, several balancing routines were applied to the data at various points in the construction process. This subsection describes where the balancing was applied and why. Manual Approach Our first approach is to review the input data and consider why cer- tain accounts do not balance. As we review each account discrepancy, 234 Part II: Detailed Analysis we often find that the largest imbalances are caused by errors in data translation or errors in data mapping. When the data are translated properly, or are mapped more tightly to the true meaning of the eco- nomic activity, many of the larger errors do not require mechanical adjustments. This has been the case for some accounts in Armenia. For example, this country has traditionally been a major center for cut diamonds. Uncut and cut gems are the single largest import and export items for Armenia. Two issues arose while generating the national accounts for precious gems. First, the national accounts were inconsistent. Total exports were US$259 million, and imports were US$214 million; but the NSS also reported that domestic production was only US$9 mil- lion, when the accounting identity implies that it is at least US$45 mil- lion. The second issue comes from the surrogate IO table. The precious stones sector is aggregated together with other mining and minerals in the GTAP database, which is the source of the IO table. Thus, the production shares for this sector did not reflect the fact that in Arme- nia more than 90 percent of intermediate demand is own-use (uncut gems). The IO table was manually adjusted to reflect expert opinion related to the jewelry industry there. Other accounts were adjusted using the least-squares methodology. The main point of this exposition is to demonstrate that "SAM balancing" cannot be a completely automatic process. As described in Round (2003, p. 174), . . . Data reconciliation methods were not as arbitrary as it might at first seem. There were essentially three steps involved in the judgement approach. First, the initial data were set alongside each other in the accounting framework to take initial stock of the problem. Secondly, a qualitative judgement was taken on the relative reliability of the alternative estimates, relying on expert local advice. Thirdly, after choosing the most reliable estimates, further scaling and adjustments were made manually to achieve consistency The same three steps are taken here, except that the third step can now be achieved using a computer. Mechanical Approach After some work using the manual approach, the remaining loose ends (and there are plenty) are balanced using a least-squares approach. Most of the new imbalances arise as we disaggregate intermediate demand (ID0ij), final demand (C0j), government demand (g0j), and A Social Accounting Matrix for Armenia 235 factor incomes (VA0Fj). After disaggregation from the macro SAM, we face the problem that some (if not all) of the basic accounting identi- ties above no longer hold, as they did for the aggregate SAM. In the least-squares approach, ID0ij, C0j, and D0j are adjusted in order to satisfy the producer competition (zero profits) and market equilibrium (supply-demand) equations. Later in the build process, we enforce the income-expenditure identities for each household decile. We describe the first problem, balancing supply and demand, first. The market-balance problem is defined in the recalibration routine as follows: min i 0 1 2 0 1 0 (8A.1.5) , j(IDi, j- IDi, j)2 + i Di - Di ( ) + i Ci - Ci ( ) 2 10 such that Mi + Di = j IDi 0 (8A.1.6) , j+ Ci + Gi + Ii 0 0 Di + Xi = j IDj 0 ,i + l L0l ,i + k Kk,i + ti 0 0 This minimization problem chooses values for ID, D, and C with zero-profit and market-balance conditions as constraints. We have applied different weights to each choice variable. The choice of weights in the objective function is somewhat arbitrary. In this prob- lem, consumption and domestic supply are weighted less heavily, so that changes in these values do not penalize the objective function as much as changes to ID. In sam.gms, another least-squares rebalancing problem is solved. Instead of squaring the absolute difference between input variables, in sam.gms we chose to square the percentage difference between input variables. For example, government consumption is adjusted accord- ing to this penalty function: Gi - 1 Gi 2 (8A.1.7) 0 i The SAM-balancing problem in sam.gms includes all of the social accounts, including taxes, international trade and savings, and gov- ernment collections and expenditures. Remaining imbalances in any account are corrected here. Finally, the household income and expenditures are fit to macro data, also by using a mechanical balancing approach. In the GAMS program file, readhh.gms, we separately fit income to aggregate factor income and transfers, and we fit expenditures to total net-domestic supply. In each case, we adjust the percentage difference between the final variables and the target values.11 236 Part II: Detailed Analysis Table 8A.1.8 Mapping of Armenian Output Statistics to SAM Sectors Total production (Y0) Armenian sector (million dram) SAM sectors (%) Industry 611,879 FOD (35.7) TBC (3.1) CRP (2.3) MNM (9.2) MCH (1.2) OMF (1.3) MIN (9.8) UTL (24.4) Agriculture 522,634 GRN (18.7) VFR (23) VOL (0.2) OCR (15.4) MIL (30.6) OMT (12.2) Forestry 942 OCR (100) Construction 297,088 CON (100) Transport & communication 142,670 TRN (100) Retail trade & catering 233,889 TRD (100) Public procurements 8,576 GOV (100) Spare parts 1,526 TRD (100) Information & tech. srvcs 1,288 OSR (100) Real estate transactions 4,914 OSR (100) General commerce 5,295 TRD (100) Geology 417 MIN (100) Other branches 3,300 OSR (100) Housing 52,471 DWE (100) Utility sector 84,605 UTL ELE Health, sport 68,277 GOV (100) Education 52,541 GOV (100) Culture 9,630 OSR (100) Science 6,988 GOV (100) Lending 26,748 BNK (100) Insurance 1,485 BNK (100) Governance and defense 71,791 GOV (100) NGOs 10,645 GOV (100) Financial intermediaries 0.0 (Not Included) Source: Armenia SAM. Note: This mapping is available as an Excel spreadsheet, included as part of the distribution archive: mapping.xls. Industrial shares were apportioned based on IMF statistics. A full listing of the IMF industrial shares is available in the program distribution: Armenia SAM. Agriculture shares were apportioned according to the Armenian national accounts, provided by the NSS, and relisted in the SAM distribution file as: Armenia agriculture 2002.xls. A Social Accounting Matrix for Armenia 237 Table 8A.1.9 Mapping GTAP V5 Sectors onto 2002 Armenian SAM Sectors (continued) SAM code Survey code Survey description MIL 10408 Cheese of cow milk (chanack, lory, other) MIL 10409 Homemade cheese, cow milk MIL 10410 Gruyere and Dutch cheeses MIL 10411 Cream cheese MIL 10412 Ice cream MIL 10499 Other milk products MIL 10501 Eggs MIL 10502 Dry eggs MIL 10601 Butter MIL 10602 Melted butter MIL 10603 Margarine VOL 10604 Sunflower oil VOL 10605 Olive oil VOL 10606 Other vegetable oils VOL 10699 All sorts of oil and animal fat VFR 10701 Apples . . . CRP 20502 Laundry detergent CRP 20503 Laundry whitening CRP 20504 Other laundry goods CRP 20599 Other cleaning materials OMF 20601 Linoleum OMF 20602 Pile OMF 20603 Wallpaper OMF 20605 Basin OMF 20606 Toilet CON 20607 Cement OMF 20609 Oil-based paint Source: Armenian Living Standards Measurement Survey (LSMS). THE BUILD DISTRIBUTION The main text describes the SAM development process, but the actual program files used to generate the dataset can appear completely differ- ent. This section of the Annex steps the reader through the program files and describes how they are combined to build a complete SAM. The SAM is based solely upon the input data files provided by the local experts. There are places in the process, however, that require manual adjustment techniques. These manual adjustments are difficult to avoid, and they are different for each country. For this reason, it is unlikely that readers will be able simply to drop in new input files and 238 Part II: Detailed Analysis mappings for their own country and expect a reasonable, balanced SAM to materialize. This distribution offers a structured approach to the thought process while constructing the SAM, but it is not a sub- stitute for careful consideration of a given country's economy. First we list the contents of the distribution file: Armenia SAM. zip:12 | run.bat | +---build | | aggr.gms | | Armenia.xls | | buildsam.gms | | extract.zip | | recalib.gms | | setup.bat | | shares.gms | | soedata.gms | | tables.gms | | | \---data | arm25.map | gtap001.set | POL.GDX | +---sam | Armenia_hh_2002.xls | hhsam.gms | printxls.gms | readhh.gms | sam.gms | \---SourceData 2002_GDP_NSS.xls Armenia_agriculture_2002.xls Export Import data in 2002 Revised.xls IMF-Statistical-Annex(Industry Shares).pdf There are three directories: build, sam, and sourceData. Most of the data management is handled in the build directory, where setup.bat directs the aggregation, mapping, and combination of each of the four input tables. The data is recalibrated using recalib.gms, the main pro- gram in the build directory. Most of the remaining GAMS programs A Social Accounting Matrix for Armenia 239 are optional; they produce descriptive tables based upon the pre- and post-calibrated data. A subdirectory, data, contains the IO table for Poland and the mapping between GTAP and our 25-sector SAM. Finally, extract.zip contains the necessary files to extract an IO table from the GTAP database. The extraction process requires the entire GTAP database, a 3MB file (gtap001.zip). The programs in the SAM directory write out the micro-SAM and then generate the household SAM. sam.gms is handed data from the build process, re-checks the calibration, then writes the micro-SAM to a comma-delimited file. Then the household dataset, Armenia hh 2002.xls, is read by readhh.gms and combined with the national accounts by hhsam.gms. printxls.gms is the program that "flat- tens" the social accounts into a single, two-dimensional matrix. Finally, a minimal SourceData directory is included. This directory contains spreadsheets that were used to build Input Tables #1 and #2, output, value added, trade, and taxation, for each sector. The most important file is run.bat, the command file. If everything is correctly in place, this file builds and writes the entire social account- ing matrix from scratch. The contents of this file are shown here: cd build call setup pol cd ..\sam call gams sam call gams printxls --dataset=ekasam call gams readhh call gams hhsam rem workbook: : call e:\gams211\xlwrite msam.csv sam.csv hhsam.csv Armenia_2002_SAM.xls if `%1'=='cleanup' goto cleanup cd .. goto end :cleanup del *.lst del *.gdx del msam.txt del *.csv cd .. goto end :end 240 Part II: Detailed Analysis First, we enter the build directory and call the setup.bat subcom- mand file, indicating that we wish to use POL for the surrogate IO table. This extracts Poland's production structure from the GTAP database13 and computes intermediate demand shares, value-added shares, and final consumption shares. The program recalib.gms com- bines the source output and tax data together with the surrogate data to generate most of the micro social accounts. These data are returned in a GAMS data container labeled sam2002.gdx. NOTES 1. This chapter was prepared by Miles K. Light, University of Colorado, and by Ekaterine Vashakmadze and Artsvi Khatchatryan, both at the World Bank. 2. While the model has limitations in many aspects, its simplicity offers certain advantages: (i) its underlying assumptions of fixed prices and fixed technology in an economy with ample underutilized capacity like Armenia are reasonable at present; (ii) the model can be applied for understanding the nature of the backward and forward linkages in the economy; (iii) the simplicity of the model permits both easy tracking of information gaps and easy transfer to the relevant counterparts for technical capacity building. At the same time, at this stage in the data development process, we advise cau- tious use of input-output analysis, since the IO table we currently use is an adjusted surrogate table. 3. Interestingly, the labor/capital intensity is reversed for high-income countries. In the United States, for example, agriculture is a capital-intensive industry and chemical manufacturing is labor-intensive, due to high R&D costs. 4. Note that US$302 million equals the ROW-SUM along the GOV accounts in the macro SAM, minus the transfers from ROW. 5. Figures are taken from Armenia (2003c), p. 116, Table 4. 6. See, for example, Robinson et al. (2001). 7. http://www.gtap.agecon.purdue.edu/. 8. http://www.gams.com/. 9. In this treatment, each producing sector provides only one output good. 10. Table 8A.1.6 contains descriptions for each symbol. 11. The target values are the original household income and expenditure figures based upon the 2002 Armenian Household Survey. 12. This archive is freely available to interested readers. It can be downloaded from the Armenia Project Web site, located here: http://www.mileslight.com/ armenia/. A Social Accounting Matrix for Armenia 241 13. This distribution does not include the full GTAP dataset. Instead, we have extracted these data and placed them into the GAMS data container, pol25.gdx. Chapter 9 Taxation and Economic Efficiency in Armenia INTRODUCTION While Armenian economic growth has been robust for the past few years, tax revenue collections have fallen short of expectations.1 Tax revenue as a fraction of GDP has fallen steadily, particularly through reductions in the collection of direct taxes. In total, tax revenues have fallen by more than 2.5 percent of GDP, with 1 percent due to reduc- tions in personal income tax collections and 1.25 percent due to reduc- tions in receipts from the profits tax. While early stages in transition and development may be led by the private sector, sustained development and reductions in poverty require public sector resources. Health, education, and public infra- structure are all essential for sustainable development, and the provi- sion of these goods and services demands an efficient and reliable tax system. There are several recent studies of the Armenian tax system that have evaluated the prospects for tax reform from legal and structural viewpoints.2 The present paper reports results from a CGE model that has been formulated to explore, from a quantitative perspective, the economic consequences of different types of tax reforms in Armenia. We use the model to consider the marginal cost of funds from direct and indirect tax instruments, and we also present a set of scenarios for increasing tax revenue by between 0.5 and 2 percent of GDP. Data discrepancies and omissions make Armenian tax policy analy- sis difficult. Presently, there does not exist a "post-Soviet" IO table or any comprehensive source of economic data for Armenia. In the absence of a consistent set of national economic statistics, our model is based on information from a variety of sources. Sectoral value added, aggregate output shares, trade statistics, and components of final 243 244 Part II: Detailed Analysis demand are based on data from the NSS. Input-output coefficients in our model are based on recent IO tables from 1997 for Hungary and Poland. We use two alternative sources of these coefficients in order to evaluate the robustness of our result. Our findings indicate that the results are not overly sensitive to the choice of production technology. The results are much more sensitive to benchmark tax rates and the elasticities of substitution that characterize trade-offs between formal and informal goods and services. Despite the data limitations inherent in this study, we still believe that policy decisions can be better informed by a roughly parameter- ized model than by the uncalibrated logic of economic theory. For example, the model can be used to assess how collections in tax rev- enue from one tax instrument are affected by changes in other tax rates. These types of assessments are particularly important in view of the high and growing level of informal activity. With some effort on the part of statistical authorities, the input data for our model can be substantially improved; thus we regard the present set of results as preliminary. The general lessons the model is now providing are, however, highly informative. The standard "first- best" rules for taxation (low rates, broad base), which apply in many competitive economies, may be inappropriate in a second-best setting. Armenia clearly represents a second-best economic environment, as is evidenced by the low collection rates of its principal taxes. Revenues from income and profits taxes are reduced by underreporting of both labor and capital income.3 VAT revenues appear to have a somewhat higher rate of compliance, but as of 2002 the relative efficiency of VAT collections is about half that of developed countries. The NSS includes an estimate of informal economic activity as part of the officially tabulated GDP. According to this estimate, one-third of the US$2.9 billion Armenian GDP is produced within the informal sector. Labor and capital income from this sector are omitted from the tax base, and tax collections based on payments within the informal sector are low. Additionally, factor earnings from informal firms can more easily evade VAT and profits taxes, or a portion of them. Several international teams, including the IMF, United States Agency for International Development (USAID), and the World Bank, have assessed the Armenian tax system and made recommendations for tax reform. Although the foci of these analyses all differ, there are some common themes that arise in these studies: The tax revenues in Armenia are low relative to developed coun- tries, and they are low even compared to most CIS countries. The IMF Fiscal Affairs Department reported 2004 tax income to be Taxation and Economic Efficiency in Armenia 245 17.7 percent of GDP (IMF 2004). Most alarming is the precipitous fall in tax collections between 1999 and 2003, during which time revenues fell from 20 percent to 17.2 percent of GDP. Most of this decline resulted from the fall in rates of collection for income and profits taxes. Official tax policy in Armenia has conformed to IMF recommen- dations. The VAT has only two rates (20 percent and 0 percent), personal income taxes have only three rates (0 percent, 10 per- cent, and 20 percent), and, where import tariffs exist, the rate is a uniform 10 percent. Tax collection and administration are reported to be weakly enforced. Despite a clear and simple tax code, collections are low. This could imply either weak administration or some forms of corruption, or both. The tax system is perceived to be unfair. Residents in Yerevan do not believe that tax revenues are used for public goods and services. Small taxpayers believe that the rich individuals do not pay their share, while wealthy taxpayers complain that tax audi- tors are unpredictable and punitive.4 As is apparent from a cursory study, the issues that surround tax policy in Armenia are wide ranging. Our analysis of these issues, however, focuses solely on economic welfare and the aggregate burden of tax collection. We implement a CGE model with which to assess the factors that determine how tax revenues respond to changes in tax rates.5 The study is macroeconomic in nature and does not consider individual firms, organizations, or any legal inter- pretation of the tax code. Our model is based on a dual economy in which there coexist formal and informal firms and markets. Firms operating in both the formal and informal sectors are assumed to maximize profits. Prices of goods and factors adjust so that supply equals demand in all markets. Informal activity in our model is calibrated to information regard- ing the level of informality in different sectors. In our central cases we adjust sectoral shares of informal activity to target the NSS estimate that 30 percent of the aggregate Armenian economy is represented by informal activities. Informal production in the model is able to evade taxes for profits wages, and value added. Our analysis demonstrates that rising levels of evasion produce a corresponding increase in the marginal cost of public funds. This implies that the cost-benefit test for publicly funded health, education, and infrastructure investments becomes increasingly more stringent as informality increases. In the present economy, an additional dol- lar of public expenditure costs between US$1.30 and US$1.60. When 246 Part II: Detailed Analysis we account for the long-run response of investment and capital stock to the perverse incentives introduced by the informal sector and the tax system, the marginal cost of funds might easily introduce a 100 percent premium on public expenditure. In view of the crucial role in economic development played by public expenditures and invest- ments, there are potential dire consequences for the long-term health of the Armenian economy if the trend toward increasing levels of informal activity cannot be reversed. Faced with the high efficiency costs of raising revenue, there are limited possibilities for raising significant revenue from any of the tax bases. In the short term, efficiency costs are smaller because capital is fixed and unable to escape, yet raising tax revenue remains costly due to the ability of firms and consumers to substitute untaxed, informal products for formal goods that are subject to tax. In the long-term model, we find that tax policy has costly impacts on capital accumulation and economic growth. The existing profits, income, and value-added taxes all tend to discriminate against investment in formal sectors, and the long-run perspective underscores the need for avoiding further discouragement of investment in these areas. We find that the long-run efficiency cost of taxes on agricultural products is low because existing profits and income taxes tend to discriminate against formal activities in manufacturing and industry. These taxes lead to underinvestment in industry, an effect that is par- tially offset by a tax on agricultural income. As has been pointed out elsewhere, taxes on agricultural activity face administrative difficul- ties because most farmers are small-holders. These farmers would be exempt from income and profits taxes in any case, if revenues and income are low. The marginal cost of public funds from any of the direct and indi- rect tax instruments are increased when a substantial fraction of the tax base is able to avoid payments, or when individuals are more willing to substitute informal goods and services for formal goods and services. These results provide strong support for tax policies that underscore the need to reduce tax evasion and informal activities organized primarily as a means of evading tax payments. The remainder of this chapter is organized as follows. The next section presents the model formulation. The third section presents some stylized facts regarding the economic structure of Armenia as represented in our dataset. The fourth section compares the relative efficiency among each tax stream and goes on to consider some tax proposals presented by the IMF in 2004. A detailed review of how the results are affected by the use of surrogate data sources for pro- duction technology is discussed in the fifth section. The last section provides conclusions. Taxation and Economic Efficiency in Armenia 247 A GENERAL EQUILIBRIUM MODEL FOR ARMENIA Our model represents Armenia as a small open economy with two types of economic activity: formal activities, which are subject to tax, and informal activities, which are untaxed. The model portrays an Arrow-Debreu economy with constant returns-to-scale and perfect competition across all modes of production. As a small, open economy, Armenia faces fixed relative prices for imports and exports. Produc- ers maximize profits taking prices as given, and consumers maximize utility subject to a budget constraint that depends upon the value of their endowments, transfers from the government, and remittances from abroad. These assumptions imply that no producer earns above- normal profits and that consumers cannot increase consumption of all goods simultaneously. These are the basic economic concepts of economic scarcity and competition. Following Mathiesen (1985), we formulate and solve the model as a complementarity problem with three types of equilibrium conditions: market clearance, zero profit, and income balance. Production tech- nology and consumer preferences are characterized using the nested, constant-elasticity of substitution (CES) functional form. The model accommodates analysis of both the static and steady-state welfare effects through alternative representations of the capital stock. The numerical equations are based on data derived from the 2002 Armenian National Accounts together with reports provided by the IMF, USAID, GTAP, and the World Bank. The present version of the model distinguishes 25 industries, the government, and a single, rep- resentative consumer. In each industry, in the reference equilibrium a given share of production is produced informally.6 Economic Flows The relationship between different sectors and consumers in the model is shown in Figure 9.1. Various aspects of the economy are depicted here, with the exception of taxes subsequently. Production in sector i (Yi) combines four primary factors: capital (K), skilled labor (LS), unskilled labor (LU), and land (N). Intermediate inputs are added to produce outputs for the domestic (Di) and export (Ei) markets. An "Armington composite good" (A) is a combination of domestic goods (D) and imports (M). Armington aggregate goods are the basic consumption commodity. They are consumed by industry as an intermediate input and they are also goods for final consumption (C), government consumption (G), and or investment (I). Consumers are endowed with factors of production (LS, LU, K, N), which are sold to industry (Y). They are also the final consumers, who use income 248 Part II: Detailed Analysis from factor sales to purchase Armington goods (C via A), to invest (I), or to create government services (G). For sectors, Yi, appearing in Figure 9.1 there are two associated activities: formal production (Yi ) and informal production (Yi ). The F I formal sector is subject to various taxes (VAT, profits, payroll, and income taxes), whereas the informal sector pays only those taxes col- lected at the border (import tariffs). These activities produce goods that are consumed in final and intermediate demand. A schematic for the formal/informal consumption activity is presented in Figure 9.2. Algebraic Formulation Our model is based on CES functions. CES functions are widely applied because they are globally regular and can be defined by their zeroth, first, and second order properties. This means that the loca- tion (price and quantity), slope (marginal rate of substitution), and curvature (or convexity) completely characterize a CES production or consumption function. This permits a simplified representation of production technology and consumer preferences. Using this general approach, the supply side of the Armenian model is as shown in Figure 9.1. We use to denote the elasticity of substitution for production inputs and to denote the elasticity of transformation for outputs.7 In the model, any choice for elasticities in each sector can be applied in order to reflect local expertise related to particular sectors. Production Functions Production Inputs Goods are produced according to a nested Leontief-Cobb Douglas technology. Intermediate inputs and aggregate value added enter at the top level: Y , vbi Yi = min min xji i j aji , mM . i (9.1) ai In this expression, xji represents intermediate inputs of good j from the domestic market and represents specialized imports for re-export in sector i.8 Value added represents a Cobb-Douglas aggregation of unskilled labor (LU), skilled labor (LS), capital (K), and land (N):9 vi = LUiLSiKi Ni , u s (9.2) in which constant returns to scale imply that F + I + + = 1. Taxation and Economic Efficiency in Armenia 249 Figure 9.1 Armenian Production Structure for Formal and Informal Activities C,G,I Ai = 4 Ei Di Mi = 4 Yi = 0 RA = 1 = 0 N K L A1 . . . Aj Source: Authors. Symbol Description: Y Goods production D Production sent to the domestic market A Armington aggregate good--this activity combines domestic production with imports to produce an Armington aggregate good for intermediate use or final demand. E Production which is exported M Imports L Labor inputs--labor is either skilled or unskilled K Capital input N Land inputs I Fixed investment demand. Combines goods from A to produce an investment good. G Government demand. Tax revenues purchase goods from A to produce the government good. C Final consumption demand. Final demand by households. Households sell labor and capital endowments to pay for final consumption. Elasticity of substitution for inputs. Elasticity of transformation for outputs. RA Representative agent. Note: Imports for re-export in the mineral products and precious stones sector have been omitted from this diagram. Production Outputs Each production sector Y produces two types of commodities: domes- tic goods Di and goods for export Ei. These goods are assumed to be imperfect substitutes, and they have a constant elasticity of transfor- mation. An algebraic formulation of this transformation function is written Yi = g(DiEi = i Di ) D 1+1/ 1/(1+1/) , (9.3) 250 Part II: Detailed Analysis where i D is the benchmark value share of domestic sales in total output for sector i, and corresponds to the model input etrndx. Imports The model adopts an Armington representation of the import demand. Armington goods, Ai, are produced by combining domestic goods with imports from the same sector. These goods are treated as imper- fect substitutes (for example, autos from Russia versus autos from Japan), with an Armington elasticity, DM, describing the degree to which these substitute in intermediate and final demand: Ai = i Mi ( M 1-1/DM + 1 - i ( M )D 1-1/DM i )1/(1-1/DM ) . (9.4) Some confusion can arise trying to distinguish between produc- tion, Yi, output (Di, Ei) and the consumption good (Ai). The Arming- ton aggregate good, Ai, combines domestic output, Di, with imports, Mi. Ai is the good used as an intermediate input and also for final demand. Trade Balance The shadow value of foreign exchange, , adjusts to clear the market for foreign exchange, a good which is "produced" with exports and consumed by imports: pi EEi + B = pi (Mi + ai M MYi) . (9.5) i i Holding all else equal, rising import demand will increase , which reflects increased demand for external currency. The exogenous param- eter B denotes a current account balance. Because this is a small, open economy, import and export prices (pŻi , pŻi ) are fixed exogenously. E M Consumption, Investment, and Government Final Consumption A single representative agent (RA) is endowed with primary factors of production: capital, labor, and resources. The RA demands final goods for consumption. Investment and government output also demand final goods, but the level is exogenously specified, while private demand is endogenously determined by utility maximizing behavior. The RA utility function is Cobb-Douglass as shown below: U (Ai = ) Ai i,i = 1 . (9.6) i i Taxation and Economic Efficiency in Armenia 251 The RA maximizes utility subject to a budget constraint: maxU (Ai ) Ai s.t. . (9.7) i piAi pK K + pLL + pN N + trn - I + B In this problem, the RA maximizes the utility function subject to a budget constraint. The Armenian budget constraint is equal to the total value of factor endowments (K, L, N), plus any transfers from the government, minus the cost of investment, plus the net current account balance. The current account balance for Armenia reflects substantial cash remittances from abroad, amounting to US$175 million, a substan- tial sum in comparison with US$683 in goods and services exports. Investment In the static formulation, investment demand is held constant at base- year levels. Investments are aggregated into a single, national invest- ment pool, then distributed among production and government sec- tors according to base-year accounts. Investment funds come from households and government. The level of investment can be altered in the steady-state formulation, as is discussed in the next section, "Armenia's Economic Structure." Government Government purchases of goods and services are supported with tax revenue, capital earnings, and net foreign exchange transfers. The model tax system and total tax revenues are described in the next section, "Armenia's Economic Structure." Steady-State Capital A major drawback of tax policy analysis in a static model is that the capital stock is fixed and unresponsive to tax-induced changes in the net rate of return. Logically, the level of investment depends upon depreciation, interest rates, and the rate of return to capital stock. Static CGE models fail to address the changes to investment and capital stock associated with changes in the tax code. We address these issues by including a steady-state model formulation. The steady-state model allows capital and investment to change in response to tax policy in a way that is consistent with a long-run analysis. The long-run equilib- rium condition links the cost of capital with the return to capital: pinv = rK (9.8) 252 Part II: Detailed Analysis This equilibrium condition in the steady-state model is associated with an equilibrating variable, , which represents the level of the capital stock. When the return to capital rises relative to the price of investment, increases to scale up investment and reflect this arbi- trage condition. Thus, in the steady-state equilibrium, adjusts invest- ment so that the cost of capital is consistent with the return to capital. This condition is equivalent to assuming "Tobin's q" is calibrated to unity in the reference equilibrium and returns to that value in the long run. Informal Market Activities We treat informal products as close but imperfect substitutes for formal goods. Consumers and firms thus distinguish between formal and informal products, and choose between these goods on the basis of relative prices. Figure 9.2 shows how formally and informally pro- duced goods are combined to produce a good "X" that is consumed across all types of demand: final demand by consumers, investment, the government, and intermediate demand by firms. x denotes the elasticity of substitution between each good type. In an economy where underground goods or services are qualitatively similar to for- mal products, or where informal activities are commonplace, there would be a high value of x. As will be shown below, as the value of x increases, so too does the cost of public funds. Figure 9.2 Demand for Formal and Informal Goods C,G,I Y X = 4 YF YI TL, TK, VAT, TM, TXS TM Source: Author. Note: Formally produced goods are subject to all official taxes, while informally produced goods prices incorporate only import tariffs. Goods from each type of production are imperfect substitutes in final and intermediate demand. Final demand and intermediate demand within each sector has a different share of informally produced goods. Taxation and Economic Efficiency in Armenia 253 Tax Structure Production inputs are subject to five major types of taxes. Final con- sumption is taxed at rate vati. Labor income in the formal economy is taxed at rate tF, and social security taxes are also imposed as a tax on labor income in the formal economy, applied at rate tpyrl. The total tax rate on labor in the formal sector is then tL = tF +tpyrl. Capital earnings in the formal economy are taxed at rate tK, imports pay tariffs at rate tm, and land rents are taxed like capital returns, where a national rate for profits taxes is applied. Differences in VAT, import, or formal/informal tax rates across sectors lead to efficiency costs that are captured in the model. An important segment of the tributary system is the invoice- rebate feature available for the VAT. Under this system, value-added taxes paid for intermediate inputs can be reclaimed by the firm. In theory, the rebate eliminates the tax distortion between intermediate inputs. This distinction is less clear in practice, because the paperwork required for collection is complicated and repayments are unpredict- able. According to a 2004 survey of Armenian companies, both of these difficulties exist. The tax-inclusive cost of production for formally-produced goods is then Ci = pjxji + pL (1 + tL L + (1 + tK ) )(rkKi + riNi ) +pi ai . M m (9.9) j Tax-inclusive revenue value for Y is denoted as Ri: Ri = pi Di + pi X . D X (9.10) In equilibrium, the tax-inclusive cost of production equals output value across all sectors (Ri = Ci), and this represents the zero-profit equilibrium condition. Import tariffs and value added taxes are included in the Armington commodity's unit cost function for formally produced goods: 1/(1-) pi = (1+ vati M )i pi (1 1- + tiM D - M pi 1 . (9.11) M D pi ) + pi The benchmark tax rate applied on formal labor inputs (tL) is based on direct tax payment by households in the SAM and gross payments to formal labor. The benchmark tax rate applied to private capital (tK) is based on the direct tax payments by private firms and the gross payments to capital in all nongovernment sectors. 254 Part II: Detailed Analysis There is perfect arbitrage in factor markets, so there is a wedge between the marginal product of labor, capital, and land in the formal and informal economies. One important aspect of the efficiency cost of taxation therefore corresponds to this difference in productivity. Any policy that leads to an increase in informal activity therefore exacer- bates this inefficiency. ARMENIA'S ECONOMIC STRUCTURE Our base-year statistics come from a variety of sources. We target the model to match official figures provided by the NSS. Although our figures do not match NSS figures to the exact dram or dollar, our benchmark represents a more consistent accounting framework than is provided by the NSS. By combining information from disparate (official) sources, and by adjusting the data to match up with the equilibrium conditions, we have been able to uncover several incon- sistencies in the Armenian national accounts. The Armenian GDP in 2002 was 1.36 trillion Armenian drams. This is equal to US$2.3 billion at a market exchange rate of 573 drams per dollar. According to the CIA's World Factbook for 2003,10 Arme- nia's GDP was US$12.13 billion in 2002, on a purchasing power parity basis. The official population, according to a 2001 Armenian census, is 3.0 million, so that annual GDP per capita in 2002 was US$766 at market exchange rates. Our model represents the economy through 25 sectors of produc- tion, trade, and consumption. The sectoral aggregation was tailored to highlight the most important industries and goods in the coun- try (subject to the availability of data). IO coefficients for the model are drawn from the 1997 IO table for Hungary, which is part of the GTAP5.4 database (Hertel 1997).11 Table 9.1 describes the sectors that are in our model, and Table 9.2 ranks these sectors by output.12 Since 1997, Armenia has enjoyed strong economic growth: over the period 2000­03, the economy grew at an annual rate of 12 percent, and inflation has remained low. Most SMEs are privatized, and the antiquated energy system inherited from the Soviet era has been mod- ernized. The country's nuclear power station, Metsamor, produces sufficient electricity, so that Armenia is now a net electricity exporter. Armenia still depends upon imported oil and natural gas, all of which comes from Iran as a result of the conflict with the energy-rich neigh- boring state of Azerbaijan. Armenia maintains a large trade deficit which has been offset by remittances, international aid, and FDI. Eco- nomic ties with Russia remain close, especially in the energy sector. Taxation and Economic Efficiency in Armenia 255 Table 9.1 Economic Sectors in the Armenian Model Code Description Y0 (US$ million) Agriculture GRN Wheat, potatoes, legumes 170.7 VFR Vegetables, fruits, grapes 209.3 VOL Vegetable oils and fats 1.4 MIL Dairy products 279.2 OMT Beef, pork, poultry 110.9 OCR Other crops 140.1 Industry ENR Energy: oil and natural gas 0 MIN Mining and quarrying 95.0 FOD Food processing & beverages 381.0 TBC Tobacco products 33.1 LMF Light manufacturing and textiles 65.7 MTL Metals and metal products 121.2 CRP Chemicals, rubbers, and plastics 105.2 MNM Mineral products and precious stones 311.7 MCH Equipment, motor vehicles, and optical 119.9 OMF Other manufacturing 77.6 UTL Electricity, gas, and water supply 146.5 Services and other CON Construction 519.8 TRD Retail & wholesale trade, catering 434.8 GOV Governance, defense, and public expenditures 364.8 TRN Transport and communications 248.9 OSR Other services 221.2 DWE Housing and dwellings 91.6 BNK Banking, lending, and insurance 87.0 Source: Totals from NSS. Individual sectors: authors' calculations. See Annex for data details. Y0 Base-year (2001) sectoral output (US$ million) In 2002, the single largest industry was construction. Growth in this sector was driven mostly by charitable donations from the Ameri- can-Armenian diaspora, such as the Lyndsey foundation. Food and agriculture represent almost 45 percent of the country's output. The remaining industry reflects the country's legacy from the Soviet era where metal-cutting machine tools, forging machines, electric motors, instruments, tires, chemicals, gem cutting, and brandy-making were the major tradable goods and services. Most heavy industries have 256 Part II: Detailed Analysis Table 9.2 Base-Year Production and Trade Statistics by Sector for Armenia, 2002 (US$ million) Code Description Y0 VA D0 E0 M0 CON Construction 519.8 297.4 513.6 6.2 2.9 TRD Retail & wholesale trade, catering 434.8 264.8 434.8 0.0 0.0 FOD Food processing & beverages 381.0 156.9 329.8 51.2 44.0 GOV Governance, defense, and public expenditures 364.8 208.8 285.8 79.0 68.5 MNM Mineral products and precious stones 311.7 40.4 52.5 259.2 212.3 MIL Dairy products 279.2 169.9 278.7 0.5 6.6 TRN Transport and communications 248.9 140.3 166.9 82.0 141.1 OSR Other services 221.2 213.6 211.3 9.9 0.8 VFR Vegetables, fruits, grapes 209.3 127.6 203.7 5.6 24.0 GRN Wheat, potatoes, legumes 170.7 104.1 170.4 0.3 50.8 UTL Electricity, gas, and water supply 146.5 101.1 140.8 5.7 13.4 OCR Other crops 140.1 85.4 135.1 5.0 19.5 MTL Metals and metal products 121.2 59.8 76.4 44.8 55.7 MCH Equipment, motor vehicles, and optical 119.9 46.3 63.6 56.3 160.3 OMT Beef, pork, poultry 110.9 65.5 110.8 0.1 21.4 CRP Chemicals, rubbers, and plastics 105.3 90.5 96.9 8.4 84.6 MIN Mining and quarrying 95.0 37.4 52.9 42.1 22.9 DWE Housing and dwellings 91.6 0.0 91.6 0.0 0.0 BNK Banking, lending, and insurance 87.0 71.8 80.3 6.7 11.3 OMF Other manufacturing 77.6 49.4 72.5 5.1 47.8 LMF Light manufacturing and textiles 65.7 22.6 36.1 29.6 40.7 TBC Tobacco products 33.1 12.7 29.5 3.6 28.6 VOL Vegetable oils and fats 1.4 0.9 1.4 0.0 16.8 ENR Oil & natural gas 0.0 0.0 0.0 0.0 151.4 Total 4,336.6 2,367.0 3,635.3 701.3 1,225.4 Source: Authors' calculations based upon total supply provided by the NSS (2002). Key: Y0 Base-year output VA Base-year value added D0 Base-year supply to domestic market E0 Base-year exports (FOB) M0 Base-year imports (CIF, net tariff) Taxation and Economic Efficiency in Armenia 257 declined precipitously since 1990. Some of these activities have been replaced by high-technology manufacturing and software services. These sectors are still small, however, and they must compete in a global market for business and ITC services. International trade statistics are presented in Tables 9.3 and 9.4. Armenia's open trade policy has been hampered by the closure of its borders with Turkey and Azerbaijan. All trade is shipped from the north through Georgia or from the south via Iran. Personal imports of goods make up an unknown but potentially significant portion of total imports. In value terms, the largest import and export is jewelry, gems, and cut stones (model sector MNM). The gem-cutting industry adds value and reexports the gems at a higher price. The key import sectors for final consumption are oil and gasoline, food and food products, and manufactures such as automobiles, machinery, and computers. Armenian exports are limited to processed gems and jewelry, preci- sion instruments, tourism and related transportation, and some gold, precious stones, and minerals.13 Sectoral value added, ranked accord- ing to labor intensity, is shown in Table 9.5. The Informal Economy A portion of the officially reported economic statistics in Armenia comes from a survey of the informal economy. This portion of eco- nomic activity is not reported to the government, but represents a cer- tain amount of economic production and consumption. Traditionally, informal economic activity is small-plot farming and marketing. Other examples are certain domestic services and street markets. The NSS estimates that in 1999 the informal economy represented 26 percent of total economic activity. The estimate for 2002 is said to have risen to 30 percent, but the official statistics are not yet available. In order to identify the nature and size of these activities, the NSS conducts a survey of 9,000 individuals. The questions they ask, or how they tabulate the value of output based upon the survey was not pro- vided. Table 9.6 shows the share of informal activity and employment as estimated by the NSS for 1999. We use these data to calibrate levels of informal activity at the sectoral level, contingent on the assumed informal share of aggregate GDP. We include the estimates from Table 9.6 in our model to identify the portion of production in each sector that escapes taxation. This portion of the economy is legally obligated to pay taxes, but does not. Presumptive and simplified taxes have been introduced in part to capture economic activity from the informal market, but the col- lections from these tax instruments are low and we have therefore omitted these taxes from the database and model. 258 Part II: Detailed Analysis Table 9.3 Benchmark Import Statistics for Armenia, 2002 Code Description M0 (US$ million) M0 (%) % M % VA MNM Mineral products and precious stones 212.3 17.3 79.8 13.0 MCH Equipment, motor vehicles, and optical 160.3 13.1 70.5 38.6 ENR Oil & natural gas 151.4 12.4 100.0 0.0 TRN Transport and communications 141.1 11.5 45.8 56.4 CRP Chemicals, rubbers, and plastics 84.6 6.9 46.5 86.0 GOV Governance, defense, and public expenditures 68.5 5.6 19.3 57.2 MTL Metals and metal products 55.7 4.5 42.2 49.4 GRN Wheat, potatoes, legumes 50.8 4.1 23.0 61.0 OMF Other manufacturing 47.8 3.9 39.5 63.7 FOD Food processing & beverages 44.0 3.6 11.7 41.2 LMF Light manufacturing and textiles 40.7 3.3 52.1 34.4 TBC Tobacco products 28.6 2.3 48.0 38.3 VFR Vegetables, fruits, grapes 24.0 2.0 10.6 61.0 MIN Mining and quarrying 22.9 1.9 30.2 39.3 OMT Beef, pork, poultry 21.4 1.7 16.0 59.1 OCR Other crops 19.5 1.6 12.6 61.0 VOL Vegetable oils and fats 16.8 1.4 84.8 61.0 UTL Electricity, gas, and water supply 13.4 1.1 8.7 69.0 BNK Banking, lending, and insurance 11.3 0.9 12.4 82.5 MIL Dairy products 6.6 0.5 2.3 60.9 CON Construction 2.9 0.2 0.6 57.2 OSR Other services 0.8 0.1 0.4 96.6 TRD Retail & wholesale trade, catering 0.0 0.0 0.0 60.9 TOTAL 1,225.4 100.0 0.0 0.0 Source: Aggregate values supplied by Armenian customs. Sectoral disaggregations are authors' calculations based upon shares from the NSS. Key: M0 Base-year imports (US$ million) M0 (%) Base-year imports as % of total imports % M Base-year imports as % of domestic sales % VA Base-year imports as a percent of sectoral value added Taxation and Economic Efficiency in Armenia 259 Table 9.4 Export Statistics for Armenia, 2002 Code Description X0 (US$ million) X0 (%) % X % VA MNM Mineral products and precious stones 259.2 37.0 83.2 13.0 TRN Transport and communications 82.0 11.7 32.9 56.4 GOV Governance, defense, and public expenditures 79.0 11.3 21.7 57.2 MCH Equipment, motor vehicles, and optical 56.3 8.0 47.0 38.6 FOD Food processing & beverages 51.2 7.3 13.4 41.2 MTL Metals and metal products 44.8 6.4 37.0 49.4 MIN Mining and quarrying 42.1 6.0 44.3 39.3 LMF Light manufacturing and textiles 29.6 4.2 45.1 34.4 OSR Other services 9.9 1.4 4.5 96.6 CRP Chemicals, rubbers, and plastics 8.4 1.2 7.9 86.0 BNK Banking, lending, and insurance 6.7 1.0 7.7 82.5 CON Construction 6.2 0.9 1.2 57.2 UTL Electricity, gas, and water supply 5.7 0.8 3.9 69.0 VFR Vegetables, fruits, grapes 5.6 0.8 2.7 61.0 OMF Other manufacturing 5.1 0.7 6.6 63.7 OCR Other crops 5.0 0.7 3.6 61.0 TBC Tobacco products 3.6 0.5 10.9 38.3 MIL Dairy products 0.5 0.1 0.2 60.9 GRN Wheat, potatoes, legumes 0.3 0.0 0.2 61.0 OMT Beef, pork, poultry 0.1 0.0 0.1 59.1 VOL Vegetable oils and fats 0.0 0.0 0.0 61.0 TRD Retail & wholesale trade, catering 0.0 0.0 0.0 60.9 TOTAL 701.3 1.0 0.0 0.0 Source: NSS (2001) (reconciled by authors). Key: X0 Base-year exports (US$ million) X0 (%) Base-year exports as % of total exports % X Base-year exports as % of domestic production % VA Base-year exports (FOB) as percentage of sectoral value added 260 Part II: Detailed Analysis Table 9.5 Sectoral Value Added, Ranked by Labor Intensity VA LAB LND SKL CAP Code Description (US$ million) (%) (%) (%) (%) MIN Mining and quarrying 37.4 72.2 0.0 9.4 18.4 LMF Light manufacturing and textiles 22.6 60.1 0.0 8.7 31.2 OMT Beef, pork, poultry 65.5 53.7 33.9 0.7 11.6 OCR Other crops 85.4 52.7 23.4 1.8 22.1 MTL Metals and metal products 59.8 52.6 0.0 8.6 38.8 GRN Wheat, potatoes, legumes 104.1 52.3 35.0 0.7 12.0 VFR Vegetables, fruits, grapes 127.6 52.2 35.0 0.8 11.9 FOD Food processing & beverages 156.9 51.2 0.0 9.4 39.4 MIL Dairy products 169.9 48.0 13.4 3.3 35.3 OMF Other manufacturing 49.4 46.0 0.0 6.9 47.0 MNM Mineral products and precious stones 40.4 42.8 0.0 6.7 50.4 TRD Retail & wholesale trade, catering 264.8 42.5 0.0 7.0 50.5 VOL Vegetable oils and fats 0.9 40.6 18.3 2.8 38.3 TBC Tobacco products 12.7 40.0 0.0 5.6 54.4 CON Construction 297.4 37.0 0.0 5.7 57.3 UTL Electricity, gas, and water supply 101.1 34.4 0.0 14.3 51.3 MCH Equipment, motor vehicles, and optical 46.3 32.6 0.0 8.6 58.8 CRP Chemicals, rubbers, and plastics 90.5 32.1 0.0 8.2 59.6 TRN Transport and communications 140.3 30.2 0.0 9.6 60.2 GOV Governance, defense, and public expenditures 208.8 25.0 0.0 39.8 35.1 BNK Banking, lending, and insurance 71.8 22.0 0.0 19.5 58.5 OSR Other services 213.6 12.4 0.0 11.0 76.6 TOTAL 2,367.0 0.0 0.0 0.0 0.0 Source: Base shares from GTAP database. Some sectors were adjusted to reflect Armenia country-office staff calculations. Key: VA Sectoral value added at factor cost (US$ million) LAB Unskilled labor share of sectoral value added (%) LND Land share of sectoral value added (%) SKL Skilled labor share of sectoral value added (%) CAP Capital share of sectoral value added (%) Taxation and Economic Efficiency in Armenia 261 Table 9.6 Estimated Level of Underground and Informal Activity Percentage Corresponding sectors in model Industry 28.7 MIN,FOD,TBC,LMF, CRP, MNM,MTL,MCH,OMF Construction 46.1 CON Transport and communications 21.1 TRN Trade 75.5 TRD Agriculture 21.0 GRN,V F,VOL,OCR,MIL,OMT Other branches 27.1 ENR,UTL,BNK,GOV,OSR,DWE GDP at market prices 28.9 Source: NSS (1999). Tax Revenue Table 9.7 shows that total government collections in 2002 were 242.3 billion drams (US$422.8 million). Value-added taxes were the larg- est revenue source, contributing 39.2 percent of total tax revenues (US$165.8 million). Excise taxes were 14.6 percent of total taxes, and payroll contributions to social security totaled 15.7 percent of revenue. These tax bases are followed distantly by enterprise taxes, income taxes, and other taxes. Table 9.7 shows tax collections for the major levies in Armenia during 2002, and Table 9.8 further disaggregates these collections by production sector. Armenia's statutory tax code is straightforward. There exists a sin- gle VAT rate of 20 percent, a single tariff rate of 10 percent, and a sin- gle low-profit tax (20 percent). Income taxes are 10 percent for 80,000 drams per month (US$139), or 20 percent if income is above 80,000 drams. Excise rates are higher, but they are only applied to tobacco products, alcoholic beverages, and petroleum. Tax collections are more complicated, but we can review the trend in collections briefly. Profits tax collections have declined precipitously since 1999, as have income tax collections. Collections from profits taxes fell by 50 percent between 1999 and 2003, from 2.2 percent of GDP to 1.1 percent. Similarly, personal income taxes fell from 1.9 per- cent to 1.0 percent over the same period. The combined loss is 2.0 percent of GDP (US$58 million if considered in 2002). The trend is strange in light of the very strong economic growth over the same period, during which time GDP increased by approximately 30 per- cent. No other tax streams have risen to compensate for this loss, and total government revenues have declined as a proportion of GDP. 262 Part II: Background Papers Table 9.7 Armenian Tax Collections, 2002 (by source) Tax stream Billion drams US$ million Percent Value-added tax 95.0 165.8 39.2 Excise tax 35.3 62.6 14.6 Profits tax 17.4 30.3 7.2 Personal income tax 12.5 21.8 5.2 Payroll taxes 37.9 66.1 15.7 Import duties 9.8 17.1 4.0 Main streams: 207.9 363.3 85.9 Other taxes (omitted from the model) Stamp taxes 14.3 24.9 5.8 Environment and property 5.4 9.4 3.7 Presumptive tax 6.4 11.2 2.6 Simplified tax 3.7 6.5 1.5 Other streams 29.8 52.0 13.6 Total 237.7 415.7 99.5 Source: Table II.1, IMF Aide-Mémoire (2004). ILLUSTRATIVE SIMULATIONS Table 9.9 presents welfare-cost estimates of raising funds from four major revenue streams. The experiments on which these calculations are based consider inframarginal changes in rates; hence we label these estimates the "Average Cost of Funds" (ACF). The ACF meas- ures the efficiency cost of raising an additional US$20 million from each of the primary tax streams. The ACF values provide a very useful input to the public policy debate, specifically related to the cost-benefit calculus of public expen- ditures. When the ACF equals 1.4, this means that US$1 of public funds costs the representative consumer US$1.4. As the ACF increases, the requisite benefit through which a public project can be justified increases, and one would expect that as the ACF exceeds 1.5, fewer public expenditures are justifiable than is the case when the ACF equals 1.2. Another consideration is the shadow economy and its role in tax revenue leakage. For some of the major taxes, it is reported that non- compliance is as high as 50 percent. The tax leakage in the shadow economy is parameterized by both the benchmark share (, shown in Table 9.10) and , the elasticity of substitution between legal goods and black-market (informal) goods of the same variety. is the econ- omy-wide share of production occurring underground. Our default assumptions are = 30 percent and = 4. The average tax leakage, and the consequent average cost of funds, rises as each parameter Table 9.8 Benchmark Tax Collections, 2002 by Production Sector (US$ millions, 2001) Description Y0 VAT TK TSS TXS TL TM Total Oil & natural gas 0 53.9 0.0 0.0 28.1 0.0 0.0 81.9 Food processing & beverages 381 17.5 1.6 3 10.0 1.6 3.0 37.1 Tobacco products 33 6.3 1.2 0 24.0 0.2 1.6 33.7 Equipment, motor vehicles, and optical 120 16.3 0.4 0 0.0 0.2 3.5 20.7 Other manufacturing 78 12.8 2.5 3 0.0 1.2 0.9 20.0 Governance, defense, and public expenditures 365 0.0 0.0 13 0.0 5.8 0.0 18.4 Retail & wholesale trade, catering 435 8.1 1.8 6 0.0 2.6 0.0 18.2 Electricity, gas, and water supply 147 4.6 8.2 2 0.0 1.1 0.0 16.2 Transport and communications 249 4.0 3.7 5 0.0 2.1 0.0 14.4 Chemicals, rubbers, and plastics 105 10.9 0.4 0 0.0 0.2 0.5 12.3 Construction 520 2.3 2.8 3 0.0 1.4 0.0 9.5 Wheat, potatoes, legumes 171 9.5 0.0 0.0 0.0 0.0 0.0 9.5 Beef, pork, poultry 111 4.2 2.1 0 0.0 0.2 1.9 8.8 Light manufacturing and textiles 66 4.7 0.5 0 0.0 0.2 1.4 7.2 Mineral products and precious stones 312 4.2 0.7 0 0.0 0.2 1.2 6.7 Housing and dwellings 92 0.0 0.5 4 0.0 1.9 0.0 6.7 Mining and quarrying 95 2.5 3.2 0 0.0 0.2 0.0 6.2 Vegetable oils and fats 1 3.9 0.0 0.0 0.0 0.0 1.8 5.6 Metals and metal products 121 3.5 0.2 1 0.0 0.5 0.0 5.4 Banking lending and insurance 87 0.0 0.7 3 0.0 1.4 0.0 5.2 Other crops 140 2.1 0.0 0.0 0.0 0.0 0.5 2.6 Diary products 279 1.4 0.2 0 0.0 0.2 0.4 2.5 Vegetables, fruits, grapes 209 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other services 221 0.0 0.0 0.0 0.0 0.0 0.0 0.0 TOTAL 172.6 30.5 46 62.1 21.1 16.7 348.8 Source: Official statistics from the Ministry of Finance, provided at the authors' request. Note: Excise taxes (TXS) have been changed. Collections reported in this table (Table 9.8) were taken from Table II.1, IMF Aide-Memoire (2004). Key: Y0: Sectoral output VAT: Value-added tax revenue TK: Profits tax revenue TSS: Payroll tax (social security) revenue TXS: Excise tax revenue TL: Wage (income) tax revenue TM: Tariff revenue TOTAL: Total tax revenues of the indicated sector 264 Part II: Detailed Analysis Table 9.9 Cost of Raising 0.5% GDP (US$15 million) in Tax Revenues: A Comparison of Tax Bases Short-run R R% t% ACF-SR ACF-LR WAGE 19.1 78 118 1.3 1.6 0 PAYROLL 41.7 36 54 1.3 1.6 0 PROFITS 27.9 54 71 1.3 5.2 ­4 TARIFF 16.7 90 129 1.4 2.5 ­1 EXCISE 62.1 24 29 1.2 1.2 0 VAT 172.6 9 11 1.2 1.8 ­1 Source: Authors' calculations. Key: R 2002 base revenue (US$ million) R Revenue required as percentage of the original tax base t Required percentage increase in tax rate ACF-SR Average cost of funds in the short run (=­EV /G) ACF-LR Average cost of funds after allowing sufficient time for capital stocks () to adjust Percentage change in Armenia's aggregate capital stock as a result of additional taxes rises. The marginal cost of funds increases more rapidly than the aver- age and is most sensitive to . Table 9.9 shows some of the "central" tax estimates by tax stream. The tax stream called WAGE denotes taxes upon labor income, typi- cally deducted from worker paychecks on a monthly basis. The PAY- ROLL tax stream represents additional payments for social security and pensions. The PROFITS tax is collections for firm profits, modeled as a tax on the return to capital. In all of the scenarios, the ACF is compared to raise an additional half percent of GDP, equal to US$15 million. In the short run, when capital and investment are fixed, labor and payroll taxes have the highest cost of funds, especially for high or . Armenia has seen increase for these two tributary systems over the past four years. Personal income taxes declined by 45 percent between 1999 and 2003, from 1.9 percent of GDP to 1.0 percent. A similar (less dramatic) trend can be seen for payroll collections (3.2 percent to 2.8 percent), and profits taxes (2.2 percent to 1.1 percent). This trend can be interpreted as a broad increase of and/or . The ACF when is high (40 percent of the economy) is 1.53 and 1.42 for labor-based taxes. Capital-based taxes have a lower ACF of 1.24. Conversely, consumption-based taxes such as the VAT and excise taxes have a much lower ACF when is large. ACF estimates in the last column of Table 9.10 for TARIFF, EXCISE, and VAT are 1.19, 1.14, and 1.30, respectively. Taxation and Economic Efficiency in Armenia 265 Table 9.10 Indices of Informality and the Average Cost of Funds = 4 = 0.3 = 8 Timeframe = 0.2 = 0.3 = 0.4 = 2 = 4 = 8 = 0.4 Short run Income tax 1.24 1.29 1.36 1.24 1.29 1.40 1.58 Payroll tax 1.24 1.29 1.36 1.24 1.29 1.40 1.58 Profits tax 1.26 1.29 1.35 1.24 1.29 1.42 1.58 Import tariff 1.35 1.38 1.42 1.34 1.38 1.47 1.57 Excise tax 1.14 1.16 1.18 1.13 1.16 1.25 1.30 Value added tax 1.15 1.17 1.21 1.15 1.17 1.22 1.28 Long run (steady state) Income tax 1.53 1.59 1.70 1.55 1.59 1.69 1.91 Payroll tax 1.53 1.59 1.70 1.55 1.59 1.69 1.91 Profits tax 5.03 5.17 5.44 4.89 5.17 5.87 6.83 Import tariff 2.41 2.46 2.54 2.38 2.46 2.66 2.88 Excise tax 1.15 1.18 1.20 1.15 1.18 1.25 1.31 Value added tax 1.76 1.80 1.87 1.77 1.80 1.87 1.99 Source: Armenian general equilibrium model calculations. Key: : A measure of the acceptance of acceptance of informal goods in place of formal goods. This elasticity reflects willingness of consumer and producers to substitute formal and informal goods and services. : Base year economy-wide share of informal activity, a measure of the extent of informality within the economy. When the shadow economy () is small, capital (profits) taxes repre- sent an attractive revenue source, with a low ACF. This all changes in the long run. In the steady-state equilibrium we calibrate the model to an assumed equalization of the cost of capital and the rate of return. To the extent that increases in taxes directly or indirectly depress the rate of return to capital or increase the cost of capital replacement, the long-run impact of tax increases will be to reduce the level of the capital stock. This neoclassical growth mechanism can substantially increase the average cost of funds, as indicated by comparing the short- and long-run results in Table 9.10. IMF Tax Reform Packages A tax reform typically involves the combined adjustment of a number of tax rates and several tax bases. Indeed, a crucial role that can be played by a general equilibrium model is to evaluate the combined 266 Part II: Detailed Analysis effect of the simultaneous adjustment of several tax instruments. This more complex approach to tax reform is required when political fea- sibility becomes an important component in tax policy. We have constructed model-based representations of several pos- sible reform packages as is indicated in Table 9.11. These scenarios are (loosely) based upon the recommendations made in the 2004 Aide- Mémoire, as prepared by the IMF Fiscal Affairs Department. Including the agriculture sector in the tax base is considered in AGR5 and AGR. TM5 and UNIF5 present two tariff reforms. UNIF5 removes within- sector exemptions, setting the tariff rate to 5 percent for all sectors in which tariffs are currently applied while retaining trade preferences with other CIS countries. TM5 is a more profound reform that applies Table 9.11 Revenue and Welfare Impacts of Selected Tax Reforms Total revenue Change in revenue % EV Timeframe Benchmark Scenario Direct Indirect NET GDP (US$) ACF Short run AGR5 172.6 178.9 6.27 ­0.05 6.22 0.2 ­6.4 1.03 AGR 219.7 236.1 16.39 0.25 16.64 0.6 ­17.1 1.03 UNIF5 16.7 24.2 7.57 ­0.22 7.35 0.3 ­7.0 0.95 TM5 16.7 49.0 32.30 ­0.32 31.98 1.2 ­29.8 0.93 TXS10 62.1 67.9 5.76 ­0.37 5.39 0.2 ­6.1 1.13 TXS 62.1 75.8 13.73 ­0.39 13.33 0.5 ­13.7 1.03 Long run (steady state) AGR5 172.6 178.9 6.27 ­0.05 6.22 0.2 ­6.7 1.07 AGR 219.7 238.4 18.77 1.48 20.25 0.7 ­15.2 0.75 UNIF5 16.7 24.2 7.58 ­0.26 7.33 0.3 ­15.1 2.07 TM5 16.7 49.2 32.50 0.59 33.10 1.2 ­46.7 1.41 TXS10 62.1 68.5 6.39 2.21 8.60 0.3 ­6.2 0.72 TXS 62.1 76.5 14.43 2.19 16.62 0.6 ­14.1 0.85 Source: Armenian CGE model calculations. Key: AGR5 Apply a 5% value-added tax on all agricultural sectors AGR Apply a 5% tax on value added, profits, and wages in the formal agricultural sector UNIF5 Move to a uniform tariff of 5%, retaining exemptions for imports from free- trade partners TM5 Apply a uniform 5% tariff on all imports TXS10 Increase excise tax rates by 10% from current levels TXS Increase excise taxes by 10% on the current base and tax domestic tobacco Taxation and Economic Efficiency in Armenia 267 a uniform levy equal to 5 percent on imports from all trading part- ners. A target value of 5 percent was used for the ad valorem rate because about 50 percent of Armenian imports come from free trade partners such as Russia. Five percent represents half of the standard 10 percent rate. As discussed, capital- and tariff-based taxes are most efficient in short time frames, but in the long run, high tariffs are detrimental to eco- nomic growth and they encourage smuggling and higher underground activity--and the ACF for import tariffs consequently increases. Among the tax policy reforms presented here, we conclude that there are several directions in which the tax system might be improved. The elimination of preferences for agriculture in the tax system strengthens revenues over the long term and does not dis- courage economic growth. As indicated in Table 9.11, revenues raised through tariffs are efficient in the short run, with ACF values of: 0.95 (UNIF5) and 0.93 (TM5). The ACF for moving to a uniform tariff is less than 1 because the tariff-reform package increases efficiency as well as revenues. Import tariffs are less attractive in the long run, where the ACF is 2.07 and 1.42, respectively. Conversely, the AGR scenario, where agricultural activity is included into all streams of the tax system (VAT, profits, and income), has a relatively low short-run ACF of 1.03, and a long-run ACF of 0.75. The long-run efficiency cost of taxes on agricultural products is low because of existing profits and income taxes, which tend to discrimi- nate against formal activities in manufacturing and industry. These taxes lead to underinvestment in industry, an effect that is partially offset by a tax on agricultural income. In the long run, by including agriculture in the tax stream like other sectors, we collect revenues, but also increase the overall efficiency of the tax system. It is pointed out in the IMF Aide-Mémoire that taxes on agricultural activity are difficult to administer, since most farmers are smallholders. These farmers would be exempt from income and profits taxes in any case, if revenues and income are low. Large agricultural firms, however, currently enjoy preferential treatment as a side effect of well-intentioned tax breaks for small farmers. SENSITIVITY ANALYSIS In a typical CGE sensitivity analysis, results are compared across a range of elasticity values. In the present application, however, the level of uncertainty is heightened by the construction of the base- year IO table. In order to develop a better understanding of the degree of uncertainty introduced through the data construction, we have 268 Part II: Detailed Analysis repeated the calculations reported above using source data from the Hungarian IO coefficients in place of the Polish coefficients. We find that our model results are remarkably robust in this dimension. The structure of the IO matrix turns out to be a third-order determinant of model results, as can be seen in the ACF-SR and ACF-LR columns in Table 9.12. Least-squares methods are used to construct the benchmark data- base for the model. When IO coefficients are drawn from different sources, this leads to slight differences in the benchmark value shares, as suggested by comparing the R and t columns in Table 9.12. Our calibration procedure holds tax revenue and sector GDP more or less consistent, but it returns somewhat different benchmark tax rates. The new benchmark tax rates remain very close to the originals, produc- ing a negligible change to the short-run and long-run cost of funds estimates. Table 9.12 Sensitivity Analysis for the Cost of Raising 0.5% of GDP R R t ACF-SR ACF-LR Hungarian IO coefficients WAGE 19.1 78 118 1.3 1.6 0 PAYROLL 41.7 36 54 1.3 1.6 0 PROFITS 27.9 54 71 1.3 5.2 ­4 TARIFF 16.7 90 129 1.4 2.5 ­1 EXCISE 62.1 24 29 1.2 1.2 0 VAT 172.6 9 11 1.2 1.8 ­1 Polish IO coefficients WAGE 19.1 78 131 1.3 1.6 0 PAYROLL 41.7 36 60 1.3 1.6 0 PROFITS 28.1 53 77 1.3 5.5 ­5 TARIFF 16.7 90 150 1.5 2.8 ­2 EXCISE 62.1 24 31 1.2 1.5 0 VAT 172.6 9 11 1.2 2.1 ­1 Source: Authors' calculations. Key: R 2002 base revenue (US$ million) R Revenue required as percentage of the original tax base t Required percentage increase in tax rate ACF-SR Average cost of funds in the short run (= ­EV /G) ACF-LR Average cost of funds after allowing sufficient time for capital stocks () to adjust. Percentage change in Armenia's aggregate capital stock as a result of additional taxes Taxation and Economic Efficiency in Armenia 269 CONCLUSIONS This paper has demonstrated the feasibility of a quantitative analysis of tax policy issues in Armenia, despite the unavailability of current IO statistics. In the presence of missing IO statistics, policy decisions are still better informed by a roughly parameterized model than by the uncalibrated logic of economic theory. The cost of public funds in Armenia, measured as the ratio of lost consumption per additional dollar of public revenue, ranges from 1.3 to over 5, depending on the tax base and the model horizon. In the short run, the cost of public funds lies between 1.3 and 2 US$ per additional dollar of funding. Our analysis highlights the efficiency cost of informal activity. The marginal costs of public funds from any of the direct and indirect tax instruments are increased when a substantial fraction of the tax base is able to avoid payments, or when individuals are more willing to substitute informal goods and services for formal goods and services. These results provide strong support for tax policies that underscore the need to reduce tax evasion and informal activities organized pri- marily as a means of evading tax payments. Our model-based analysis emphasizes the important impact of tax policy on capital accumulation and economic growth. The exist- ing profits, income, and value-added taxes all tend to discriminate against investment in formal sectors, and the long-run perspective underscores the need for avoiding further discouragement of invest- ment in these areas. We have shown that certain "first-best" rules for taxation (low rates, broad base), which apply in many competitive economies, may be misleading in Armenia's second-best environment. ANNEX 9.1 SECTORAL CLASSIFICATIONS T he Armenian CGE model has 25 sectors. This aggregation has been chosen because it offers a reasonable characterization of the Armenian economy given data available from the Armenian Ministry of Finance and NSS, along with an ad hoc set of reports provided by various international organizations. Some aspects of the current national accounts are notable. The first is that 10 years after separating from the Soviet Union and undertak- ing market reforms, the NSS remains loyal to the original socialist accounting system. In this system, several superfluous accounts that comprise less than 0.5 percent of GDP are distinguished, whereas several more impor- tant categories of economic activity such as manufacturing and food processing are ignored. These shortcomings are rectified to a certain extent by leveraging evidence from IMF country reports, auxiliary data, and outside accounts, such as the GTAP database. This annex is designed to be comprehensive so that the inter- ested reader can recreate, append, or improve our Armenian dataset, although the NSS intends to switch accounting methods to the Inter- national Standard Industrial Classification (ISIC). SECTORAL MAPPING Some of the sectors in our accounts have been disaggregated from the Armenian national statistics. Other sectors reflect aggregations of inconsequential sectors from the accounts. A detailed description of each sector is included here. The tables used to compute total output for these sectors are provided at the end of this annex. 270 Taxation and Economic Efficiency in Armenia 271 grn Grains and legumes. This sector comes from the NSS agri- culture accounts dataset for 2002 (see Figure 9A.1.1), combining wheat, potatoes, and leguminous plants. vfr Vegetables and fruits, including grapes. This sector includes vegetables, fruit, and grapes. vol Vegetable oils. This sector is taken directly from 2002 agricul- ture accounts. sgr Sugar. This sector is taken directly from 2002 agriculture accounts. mil Milk and milk products. This sector comprises eggs and milk (without butter). omt Other meats. This sector includes the beef, pork, mutton, goat, and poultry categories from the 2002 agriculture dataset. min Mining and quarrying. This sector comes from the national accounts' (Macro Data.xls) "Geology" sector. It is also based on the IMF Statistical Annex (IMF 2002), page 7, Table 4: Structure of Industrial Production (1996­2001). The portion for "Mining and Quarrying" is taken from the "industry" sector of the national accounts. fod, tbc, crp, omf Each of these sectors is disaggregated directly from the industry macro-sector via Table 4 from the IMF Statistical Annex (IMF 2002). The shares listed for year 2001 are used. lmf Light manufacturing and textiles. This sector includes textiles, dressing, and dying of fur. mnm Minerals and mining goods. This sector mainly represents the gem-cutting business in Armenia. The single largest import and export good in Armenia is gems. Uncut gems enter the country and finished jewelry is exported. Shares are taken from Table 4 of the statistical annex. See Figure 9A.1.2 in this annex. mtl Basic metals and fabricated metals. This sector is taken directly from the IMF Statistical Annex (IMF 2002). It represents 8 percent of industrial output. mch Manufacturing. This sector combines machinery, equipment, and motor vehicles with optical, medical, and other precision devices. The combined share is of these activities in GDP is 3.8 percent. trd Wholesale and retail trade and commerce. This sector com- bines macro lines for retail trade and catering with general commerce. bnk Banking and insurance. This combines two lines from the 2002 national accounts. enr Energy This sector accounts for natural gas and oil. Energy imports are used to calculate total domestic supply. 272 Part II: Detailed Analysis utl Gas, water, and electricity. This sector appears twice; first in the industrial section of the IMF report, then again as a sepa- rate account in the official NSS accounts. We combine these two accounts for the total. The totals are the "utility sector" from the macro report and "electricity, gas, and water" from IMF (2002), Table 4. gov Government, defense, and public procurement. This sector rep- resents government activities within the economy. It is based on the macro data spreadsheet and also is compared with gen- eral statistics regarding the government sector. This sector also includes "social spending," including four lines from the national accounts: health and sport, education, culture, and science. dwe Dwellings and housing. This sector is based on the housing entry in the national accounts. con Construction. This sector is taken directly from the official NSS National Accounts (see Figure 9A.1.3 in this annex). osr Other services. This sector captures the sectors real estate, culture, and information technology. cgd Savings good. This good represents net savings and capital investment for the country. The sector consumes mostly durable goods and construction. ACCOUNTING IDENTITIES The economic model for Armenia represents an Arrow-Debreu equi- librium. This means that firms are assumed to maximize profits and face competition, households maximize utility, and markets for goods are balanced (supply equals demand). These basic conditions require a handful of accounting identities to hold. These identities are discussed in this section. First, we take the set I to represent sectors in the model {grn, min, fod, etc. . .}. Then we have the following conditions: The total sales value must equal the total cost of production: Yi =IDji + Li + Ki +Ti = Ci. (9A.1.1) j Total output or sales for good i (Yi), at producer prices, must be large enough to cover the cost of production (Ci). This includes the purchase of intermediate inputs (IDji), value added (Li, Ki), and taxes (Ti). Supply must equal demand for all markets: Ai IDij +Gi + FDi + INV , (9A.1.2) j Taxation and Economic Efficiency in Armenia 273 where Ai = Di + Mi . (9A.1.3) Total supply is represented by the "Armington Composite Good" (Ai), which combines domestic production (Di) with imported goods (Mi). Domestic endowments of value added (fi) can also be included in this condition, where fi represents the aggregated endowments of labor (Li) and capital (Ki). Domestic supply is either consumed or exported: Yi = Di + Xi . (9A.1.4) Total production for a given good (Yi) is either sold to the domestic market, or it is exported. The export share represents the value of exports divided by total output xi/yi. These basic conditions are sufficient to identify or compute most of the national accounts for Armenia. The origin of each variable appearing in the equilibrium bench- mark identities is listed in Table 9A.1.1. INPUT-OUTPUT DATA When the total production, value added, and intermediate demand values are calculated from the national accounts data, we then decom- pose the intermediate demand structure using our surrogate IO table. We follow a similar procedure to compute the capital/labor shares in value added. The current implementation distinguishes four factors: skilled labor, unskilled labor, capital, and land. Each factor's share in pro- duction is presented in the main text of this chapter. We include the dollar payments to those factors below, in the IO table. We also include payments made by producers to various tax authorities. Although the employer does not distinguish tax payments from labor costs (they are still factor payments), we know that the labor costs will change when tax rates are altered, so they are distinguished here. External Data Sources In this section, we include tables from the data sources mentioned above. Each table represents an Excel (XLS) worksheet. These data are available upon request, but they are also printed in this section for convenience. Different accounts data are stored as a set of spreadsheets and archived on the Armenia project Web site. These files have been posted to http://mileslight.com/armenia/. A password may be nec- essary to access this site. Please contact the authors if necessary. 274 Part II: Detailed Analysis Table 9A.1.1 Origin of Each Variable Yi Total supply. This parameter is given by the NSS National Accounts data, the subtables from the IMF Statistical Annex (2002), and the Agriculture subtable. Xi, Mi Imports and exports of goods and services. Provided for 2002 by the customs authorities and the NSS. Each traded good was mapped onto the model sectors. Di Domestic supply. Computed parameter based on the values for Yi and Xi. Ai Armington aggregate supply. Ai represents the total supply for a good in Armenia. It is a computed parameter, which combines Di and Mi, net of import, and value-added levies applied at the border and domestically. IDji Intermediate demand. Aggregate values for intermediate demand are provided in the national accounts data (Figure 9.5). But individual input coefficients are taken from a surrogate IO table for Hungary. This IO table is reproduced for the reader in Tables 9.1 and 9.2. Li, Ki Labor and capital demand for production in sector i. Total value added is provided in the national accounts, but the share of labor, capital, land, and indirect taxes is not provided by the NSS. These values are based upon the surrogate IO tables from the GTAP database. The capital intensity can be an important determinant of labor/capital returns, and should be reviewed more carefully by experts in Armenia. fi Factor endowments. Factor endowments for labor, capital, and land are computed from total demand by firms and government. Total sales of labor and capital are allocated to households. A data discrepancy exists between the official wage statistics, provided by the NSS, and the official value- added statistics. Although total value added is reported to be more than US$2,100 million, officially reported wages are only (approximately) US$600 million, which implies that labor's share in value added is less than 30 percent of total value added. Gi Government demand for good i. These values are based upon the GTAP surrogate dataset, but can also be reconciled by reports by the Armenian Ministry of Finance. INVi Investment demand for good i. Taken from the surrogate dataset from GTAP. FDi Final demand for good i. Computed as a residual based upon total supply and total demand. Total final demand should be compared with the surrogate data, as well as with Armenian household surveys. This task is forthcoming in a future study. Taxation and Economic Efficiency in Armenia 275 Table 9A.1.2 External Data Sources File Description armdata.xls Macroeconomic production statistics for import into GAMS. This spreadsheet was taken from Macro Data.xls. The spreadsheet reports value added only, not the total production value. Nevertheless, the shares from this sheet for year 2002 are used to calculate the share of total production in Armenia. arm ag 2002.xls Contains major agricultural sectors, including production, imports, exports, and final demand. This is a very useful spreadsheet. The agricultural information from this sheet is used in the model. Stat. Annex 2001 IMF macroeconomic statistics for Armenia for 2001 (IMF 2002). Several tables are used from this report, including industrial production (p. 7) and the statutory tax code (pp. 52­57). 276 Part Table 9A.1.3 Production Technology Structure for Armenia (US$ million) II: Sector GRN VFR VOL OCR MIL OMT MIN FOD TBC LMF CRP MNM MTL MCH OMF UTL CON TRN TRD BNK GOV OSR DWE Detailed GRN 24.35 0.01 0.00 0.16 5.50 6.93 0.00 11.82 1.21 0.03 0.01 0.56 0.02 0.01 0.02 0.00 0.43 0.03 1.02 0.00 0.39 0.01 0.00 VFR 0.00 23.61 0.00 0.10 3.98 1.53 0.20 15.94 0.88 0.06 0.06 1.07 0.13 0.05 0.15 0.00 0.91 0.13 2.89 0.05 0.75 0.05 0.00 VOL 0.60 1.04 0.31 0.49 1.48 1.75 0.00 8.82 0.03 0.10 0.01 0.09 0.01 0.00 0.03 0.00 0.05 0.00 0.27 0.00 0.16 0.00 0.00 Analysis OCR 0.30 0.89 0.01 25.35 3.49 2.20 0.34 6.78 2.66 1.73 0.03 0.41 0.05 0.04 0.28 0.01 1.18 0.15 0.94 0.00 0.25 0.01 0.00 MIL 0.48 0.42 0.00 0.30 43.59 1.05 0.31 0.56 0.24 0.08 0.03 1.18 0.12 0.04 0.04 0.02 1.16 0.13 2.91 0.05 2.74 0.05 0.00 OMT 0.19 0.00 0.11 0.18 9.14 8.85 0.00 65.46 0.12 0.07 0.00 0.00 0.00 0.00 0.00 0.00 0.41 0.00 0.84 0.00 0.34 0.01 0.00 ENR 0.02 0.27 0.00 0.05 0.42 0.05 1.03 0.94 0.15 0.06 0.52 16.58 1.32 0.01 7.87 7.95 0.17 0.03 0.00 0.00 1.77 0.06 0.00 MIN 0.00 0.06 0.00 0.00 0.00 0.00 0.78 0.05 0.01 0.02 0.39 5.43 0.45 0.01 0.03 0.44 0.84 0.08 0.03 0.00 0.19 0.01 0.00 FOD 2.57 7.44 0.02 2.71 11.00 10.64 0.31 36.34 1.37 0.45 0.08 0.71 0.21 0.07 0.14 0.01 0.45 0.21 3.03 0.00 5.35 0.06 0.00 TBC 0.00 0.00 0.00 0.13 0.00 0.00 0.00 0.29 1.90 0.01 0.02 0.00 0.00 0.00 0.02 0.00 0.00 0.07 0.26 0.00 0.59 0.07 0.00 LMF 0.53 0.82 0.00 0.73 0.79 0.08 0.59 6.33 0.62 11.96 0.45 9.22 0.40 1.07 4.17 0.07 4.68 1.30 3.57 0.33 6.37 0.22 0.00 CRP 11.03 13.04 0.02 5.23 6.37 2.30 5.12 12.07 1.46 3.57 5.96 34.55 5.50 3.62 1.63 3.99 10.05 8.92 4.29 0.11 13.17 0.23 0.00 MNM 0.85 0.99 0.00 0.50 0.53 0.24 0.83 0.91 0.58 0.24 0.18 27.77 1.67 1.00 0.12 0.13 39.45 0.39 0.97 0.01 1.19 0.25 0.00 MTL 2.29 2.55 0.00 1.28 1.63 0.63 5.69 4.05 0.72 0.67 0.50 24.64 30.04 9.30 0.46 0.48 24.33 2.04 2.99 0.20 2.21 0.24 0.00 MCH 4.30 3.90 0.01 2.53 3.36 1.21 9.25 5.99 0.82 1.24 0.65 16.83 3.71 43.61 0.51 1.73 16.59 9.78 7.19 2.04 11.59 0.35 0.00 OMF 2.49 3.52 0.01 1.50 2.27 0.50 1.69 6.08 1.36 12.10 1.63 19.89 0.69 1.64 5.16 5.01 7.15 7.03 9.31 0.54 12.76 0.43 0.00 UTL 1.29 2.24 0.00 1.63 2.20 0.70 12.36 4.79 0.45 0.99 0.83 22.66 3.44 0.64 0.36 13.07 1.27 2.83 1.74 0.25 8.11 0.33 0.00 CON 0.00 0.17 0.00 0.06 0.16 0.00 0.38 0.27 0.07 0.00 0.05 1.10 0.16 0.13 0.01 0.23 3.03 3.09 2.05 0.44 0.58 0.33 0.00 TRN 2.25 3.00 0.01 2.41 1.83 0.57 3.75 6.51 0.78 2.28 0.67 18.31 3.55 2.23 0.86 1.34 38.42 16.97 40.88 2.08 15.09 0.68 0.00 TRD 8.71 12.51 0.02 5.85 6.96 3.11 3.84 13.20 1.16 2.90 0.71 19.34 4.88 3.48 1.29 0.46 27.25 14.94 18.60 1.74 18.20 0.91 0.00 BNK 1.75 2.10 0.00 0.96 1.09 0.31 1.28 2.61 0.55 0.51 0.30 4.98 0.71 0.72 0.22 0.61 5.10 3.50 8.48 0.90 2.53 0.15 24.70 Table 9A.1.3 (continued) Sector GRN VFR VOL OCR MIL OMT MIN FOD TBC LMF CRP MNM MTL MCH OMF UTL CON TRN TRD BNK GOV OSR DWE GOV 1.11 1.30 0.00 0.91 1.33 0.45 2.73 2.32 0.34 0.66 0.25 13.34 1.22 1.02 0.32 0.29 8.59 3.95 10.03 1.06 12.26 0.58 0.00 Taxation OSR 1.49 1.75 0.00 1.57 1.78 0.19 4.62 8.90 1.72 2.69 0.91 31.80 2.38 4.36 1.08 1.36 26.64 27.42 43.39 3.42 33.79 2.55 0.00 DWE 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.03 0.00 0.02 0.00 0.02 0.00 0.00 Skilled and labor 0.76 1.01 0.02 1.51 5.69 0.46 2.95 14.20 0.71 1.96 7.46 2.72 5.16 3.99 3.44 14.42 16.96 13.45 18.59 14.00 83.20 23.48 0.00 Unskilled Economic labor 54.45 66.65 0.35 43.67 81.54 35.21 22.52 77.65 5.06 13.58 29.05 17.30 31.44 15.11 22.77 34.78 110.19 42.34 112.46 15.80 52.30 26.48 0.00 Land 36.45 44.73 0.16 21.93 22.76 22.22 5.31 5.14 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Capital 12.49 15.24 0.33 18.33 59.55 7.17 5.60 56.50 6.42 6.68 53.59 19.96 22.07 26.84 20.64 49.01 167.26 80.07 128.05 38.90 60.50 163.60 66.90 Capital tax 0.00 0.00 0.00 0.00 0.17 2.05 2.94 1.47 1.14 0.50 0.35 0.69 0.17 0.35 2.13 7.79 2.73 3.34 1.65 0.65 0.00 0.00 0.00 VAT rebate ­3.03 0.00 ­0.07 ­1.03 ­0.81 ­1.63 ­6.96 ­25.49 ­2.84 ­5.68 ­0.83 0.00 ­2.92 ­7.43 ­2.95 ­1.45 ­1.65 ­1.57 ­7.43 0.00 0.00 0.00 0.00 Effi Labor tax 0.00 0.00 0.00 0.00 0.18 0.18 0.18 1.58 0.18 0.18 0.18 0.18 0.53 0.18 1.23 1.05 1.40 2.11 2.63 1.40 5.79 0.00 0.00 Payroll tax 0.00 0.00 0.00 0.00 0.38 0.38 0.38 3.44 0.38 0.38 0.38 0.38 1.15 0.38 2.67 2.29 3.06 4.59 5.73 3.06 12.61 0.00 0.00 ciency Output 167.7 209.3 1.3 139.0 278.4 109.3 88.0 355.5 30.2 60.0 104.4 311.7 118.3 112.5 74.7 145.1 518.2 247.3 427.4 87.0 364.8 221.2 91.6 Source: Authors' calculations based on Armenia SAM. in Note: This table displays the zero profit accounting for each production sector. The table includes intermediate inputs, factor inputs, and tax Armenia payments by producers. 277 Figure 9A.1.1 Armenian Agricultural Accounts for 2002 278 min. U.S. $ Average Consumption per capita retail Stocks Food Stocks kg/ gram/ kilo- Part prices beginning Gross consump- Animal Other end of Satisfaction year day calorie/ AMD/kg of year Production Import supply tion feed Losses Seed Export use year Use rate % day II: Wheat 95.4 9.0 47.4 57.9 114.2 75.5 13.4 1.7 6.0 0.3 0.0 17.3 114.2 45.2 151.2 414.2 1325.6 Detailed Potato 154.3 20.5 100.7 0.5 121.7 32.5 19.2 7.3 32.8 0.0 0.0 29.9 121.7 99.5 40.2 110.1 82.5 Vegetables 80.0 3.1 77.5 1.0 81.6 66.2 1.6 3.5 0.4 3.1 0.0 6.7 81.6 102.8 158.1 433.2 130.0 Analysis Fruit (without grapes) 250.0 2.6 36.0 4.1 42.7 34.8 0.0 3.4 0.0 2.0 0.0 2.4 42.7 94.6 26.6 72.9 109.4 Leguminous plants 475.0 0.7 2.9 3.3 6.9 5.7 0.0 0.3 0.2 0.0 0.0 0.7 6.9 46.7 2.3 6.3 18.2 Vegetable oil 574.0 1.1 1.2 12.4 14.7 13.4 0.0 0.2 0.0 0.0 0.0 1.1 14.7 8.8 4.5 12.2 80.7 Sugar 257.9 1.4 0.0 30.8 32.2 30.4 0.0 0.5 0.0 0.0 0.0 1.3 32.2 0.0 22.5 61.6 237.8 Eggs 718.5 0.4 33.0 0.1 33.5 29.2 0.0 1.8 1.1 1.0 0.0 0.4 33.5 102.7 7.8 21.3 11.3 Milk (without butter) 250.1 20.0 213.5 4.6 238.2 189.5 21.4 9.5 0.0 0.3 0.0 17.4 238.2 98.0 144.7 396.4 158.5 Beef 1111.7 1.6 57.4 17.1 76.0 73.3 0.0 1.0 0.0 0.2 0.0 1.6 76.0 77.3 12.6 34.5 62.1 Pork 1197.1 0.4 21.5 7.9 29.9 28.8 0.0 0.4 0.0 0.0 0.0 0.6 29.9 73.1 4.6 12.6 26.4 Mutton & goat 999.5 0.2 10.3 0.0 10.5 10.1 0.0 0.2 0.0 0.0 0.0 0.2 10.5 100.0 1.9 5.3 8.5 Poultry meat 1146.7 0.8 8.8 25.6 35.2 34.2 0.0 0.2 0.0 0.0 0.0 0.8 35.2 25.6 5.7 15.6 16.7 Grape 387.0 1.5 70.2 2.4 74.1 68.5 0.0 3.4 0.0 0.4 0.0 1.8 74.1 97.2 33.8 92.6 4.2 Total 63.1 680.4 167.7 911.3 692.2 55.6 33.3 40.5 7.5 0.0 82.2 911.3 2272.0 average ex. rate 1 US$ = 573.35 AMD Note: These accounts were used to compute production values for agricultural sectors in the current Armenia dataset. Taxation and Economic Efficiency in Armenia 279 Figure 9A.1.2 Industrial Production Shares According to the National Statistical Service Total industry 100.0 100.0 100.0 100.0 100.0 100.0 Mining and quarrying 3.6 3.5 3.9 3.9 5.3 6.6 Manufacturing 74.4 69.6 62.6 64.1 64.5 64.8 Food processing and beverages 44.1 45.6 40.7 39.1 36.6 35.6 Tobacco products 0.1 0.9 2.2 3.9 3.8 2.8 Textiles, dressing, and dyeing of fur 1.7 1.5 1.7 1.3 1.3 1.4 Chemical, chem. products, rubber, plastics 4.2 3.3 2.8 2.9 3.4 3.0 Other nonmetallic mineral products 3.1 3.9 3.8 3.3 2.4 2.7 Basic metals and fabricated metal 2.3 1.5 2.6 3.6 6.6 8.0 Machinery, equipment, and motor vehicles 8.0 4.7 3.1 1.9 2.5 3.1 Medical, precision, optical instruments 1.0 1.1 0.8 0.3 0.4 0.7 Jewelry and related articles 9.3 6.0 4.0 6.3 5.7 5.4 Other 0.6 1.3 0.9 1.4 1.8 2.1 Electricity, gas, and water supply 21.9 26.9 33.5 32.0 30.2 28.6 Source: Armenian authorities, National Statistical Service. Note: Taken from IMF (2002). 280 Part II: Detailed Analysis Figure 9A.1.3 Official Armenian National Production Statistics for 2002 2002 Current prices VA IDO YD millions of drams Industry 256,925 354,955 611,879 Agriculture 318,773 203,861 522,634 Forestry 713 229 942 Construction 172,238 124,850 297,088 Transport and communication 83,640 59,031 142,670 Retail trade and public catering 144,051 89,838 233,889 Public procurements 5,703 2,873 8,576 Sear-parts 1,016 510 1,526 Information-technological provision 959 330 1,288 Real estate transactions 3,987 927 4,914 General commerce 3,538 1,758 5,295 Geology 279 137 417 Other branches 2,200 1,099 3,300 Housing 38,331 14,140 52,471 Utility sector 63,270 21,334 84,605 Health, sport 32,418 35,860 68,277 Education 43,296 9,245 52,541 Culture 6,532 3,098 9,630 Science 4,856 2,132 6,988 Lending 19,470 7,278 26,748 Insurance 1,223 261 1,485 Governance and defense 35,833 35,958 71,791 NGOs 6,283 4,362 10,645 Financial intermediaries -14,937 14,937 0 Total by branches, in main prices 1,230,597 989,003 2,219,600 Net taxes on production and import minus subsidies 131,875 Taxes 140,145 Subsidies 8,271 Gross domestic product at market price 1,362,472 Source: NSS. Taxation and Economic Efficiency in Armenia 281 NOTES 1. This paper was prepared by Miles K. Light, University of Colorado, and Thomas F. Rutherford, Ann Arbor, MI. 2. These studies include the "Tax White Paper," a survey on business opin- ions produced by the Armenian Chamber of Commerce, 2004, Chamber of Commerce and the IMF Aide-Mémoire (2004). 3. Four of the largest firms in the country reported losses for two straight years, despite GDP growth of 15 percent. 4. See the "Tax White Paper," a survey on business opinions produced by the Armenian Chamber of Commerce in 2004. 5. We do not deal with the distributional consequences of tax reform in the present analysis, although the extension of the present analysis to account for the impact of taxes on poverty would be quite interesting and surely important. 6. Our definition of the "informal market" in this analysis includes those economic activities that do not pay taxes. Part of this group is small farmers and businesses that do not have the capacity to calculate and pay their taxes; another part is those individuals or corporate entities who are explicitly evad- ing government taxes by not reporting their activities. The latter represents illegal economic activities that do not generate taxable transactions. 7. The values for and in the central scenarios are shown in Figure 3.1 8. The greatest sectoral imports in the base-year data are for sector MNM (mineral products and precious stones), in which total imports equal US$230 million. The MNM sector also generates the largest level of exports, equal to US$269 million. We characterize imports in this sector as specialized interme- diate inputs to the production sector through technology parameter ai M rather than as part of final demand. For all other sectors aiM = 0. 9. The numerical model permits the more general CES functional form for value-added based on model input esubkl. When this input is unity, value- added aggregates are Cobb-Douglas as shown here. 10. https://www.cia.gov/cia/publications/factbook/index.html. 11. The Global Trade and Analysis Project (GTAP) develops a consist- ent and balanced trade database that includes 57 production sectors and 55 regions. For more information about the GTAP database, see http://www. gtap.org. 12. A more detailed description of the sectoral classification is included in the first section of Annex 9.1. 13. While there exists a perception among Armenians that software is a major export industry, this claim is not supported by data from the NSS. The NSS reports that computer-related services account for less than 1 percent of GDP. The statistic is believed to be low because most multinational cor- porations that purchase IT services in Armenia account for the business as 282 Part II: Detailed Analysis a cost center for the corporation. Lycos is one example. The company hired 60 programmers in 2004 at an average monthly salary of US$600. According to informal sources, computer and information service exports were US$9.9 million in 2002. We have not yet been able to introduce this sector in our database and model for lack of official statistics. Chapter 10 Strengthening Competition INTRODUCTION The Armenian economy has registered sustained, rapid growth over the past decade, especially in the last few years.1 In 2003, GDP increased by 14 percent in real terms after having increased by 12.9 percent in 2002. Economic growth continued at 10.1 percent in 2004, with agriculture, construction, and services leading the growth; and at 13 percent per annum in 2005 and 2006. This growth performance is one of best in the world. The government has committed itself to eco- nomic reform and the economy is gradually being transformed. It has made private sector development the main pillar of its growth strat- egy and has been working to improve the business environment by reducing regulations, improving the bankruptcy law, improving cus- toms administration, strengthening the banking system, and reducing the capacity of officials to hamper businesses. In this process, the gov- ernment has received substantial support from donors. Nevertheless, per capita GDP remains low. Although extreme poverty and poverty have fallen strongly over recent years, there are still large numbers of poor in Armenia, especially in the rural areas and urban centers out- side the capital, Yerevan. Unemployment statistics still show a large number of unemployed, although the informal sector, which currently is about half as large as official GDP, probably has absorbed a substan- tial number of those who are recorded as being out of work. Although the growth performance of the economy has been strong, the government is anxious to ensure that it is sustainable. In this regard there is concern that the lack of competition within Armenia could threaten the long-term sustainability of growth and that it has been a contributing factor to the apparent failure of the recent eco- nomic expansion to benefit more of the population. The aim of this chapter is to arrive at a preliminary judgment on the importance of competition for the development of the Armenian 283 284 Part II: Detailed Analysis economy. The chapter examines the scope of anticompetitive behav- ior, the sectors that are the most affected, the relevant work that has been undertaken regarding these issues, the policy options that are available to deal with the problems, and the additional work that is needed. The discussion touches on the political economy issues related to areas of the economy in which competition is lacking and suggests policy options for the government. The chapter is organized as follows. We look at three main areas in which competition issues are important in Armenia. In the first section, we discuss briefly those sectors that are traditionally regu- lated and where there are natural monopolies or where competition is inherently limited, such as utilities, telecommunications, and trans- portation. In the second section we look at sectors of the economy that in Armenia are generally viewed as being controlled by a small number of incumbents who either singly or in cartels dominate the distribution of particular products. These include gasoline, wheat, cut flowers, and sugar. We also discuss some of the methods that appear to be used to maintain market dominance and make suggestions for dealing with this situation. In the third section, we look at barriers to entry into economic activities more generally. The higher the entry barriers, the less likely growth will trickle down to the population and the greater the long-run constraint on growth. The final section of the paper provides some conclusions and policy implications. ARMENIAN BUSINESS COMPETITIVENESS INDICATORS Competitiveness indicators aim to distinguish the strengths and weak- nesses of the economy and thereby identify actions to improve com- petition. The most notable institutions that produce international com- petitiveness rankings are the World Bank, the World Economic Forum (WEF), and the International Institute for Management Development. Although these research organizations include a significant number of countries in their competitiveness assessments, currently none offers a sector-specific country analysis2 that would be useful for this study. For Armenia, we used the following indicators to assess the com- petitive environment (see Tables 10.1­10.4):3 Business formation Entry-exit turnover Gross value added (GVA)4 Employment Average earnings Table 10.1 Business Formation Indicators, 2001­04 Number of registrations as a proportion of the total Number of Number of number of registrations registered entities deregistered entities at the start of the year Sector 2001 2002 2003 2004 2001 2002 2003 2004 2001 2002 2003 2004 Agriculture, forestry, and fishery 121 119 97 61 38 31 33 51 4.9 4.7 3.7 2.3 Industry (including mining) 632 443 371 252 139 141 146 203 7.7 5.1 4.1 2.7 Electricity, gas, and water supply 15 18 21 23 3 11 1 13 6.2 7.5 8.5 8.6 Str Construction 178 147 162 144 14 28 51 105 7.6 5.9 6.2 5.3 engthening Wholesale and retail trade 915 742 621 432 333 351 362 541 4.1 3.2 2.7 1.8 Transportation and communications 189 180 124 96 31 18 31 52 13.8 11.8 7.3 5.4 Financial services 23 13 15 17 4 4 1 6 13.4 6.8 7.5 7.9 Real estate, renting, and commercial activity 256 214 152 118 55 42 51 76 10.6 8.2 5.5 4.1 Education 58 35 33 28 16 6 20 17 11.4 6.4 5.7 4.7 Competition Health 62 89 53 37 70 56 24 132 7.1 10.3 5.9 4.0 Social and other services 325 204 147 113 41 49 49 97 10.5 6.0 4.2 3.1 Source: NSS. 285 286 Part II: Detailed Analysis Table 10.2 Employment Indicators, 2001­04 (%) Sector 2001 2002 2003 2004 Agriculture, forestry, and fishery 18.7 22.3 24.7 28.9 Industry (including mining) 12.3 9.7 11.1 10.3 Electricity, gas, and water supply 5.1 3.5 3.2 3.8 Construction 3.9 5.3 6.2 8.9 Wholesale and retail trade 12.5 14.1 17.4 12.6 Transportation and communications 5.1 5.5 5 4.7 Financial services 0.7 0.7 0.4 0.4 Real estate, renting, and commercial activity 1.4 0.4 1 2 Education 16.4 15.5 13.6 10.4 Health 6.8 6.5 4.2 5.5 Social and other services 5.2 5.5 5.1 5.3 Source: NSS. Table 10.3 Average Earnings Indicators, 2001­04 2001 2002 2003 2004 Agriculture, forestry, and fishery 3.0 3.1 4.1 5.6 Industry (including mining) 7.1 7.1 10.5 11.7 Electricity, gas, and water supply 11.2 11.4 12.5 14.7 Construction 8.1 7.8 9.7 11.8 Wholesale and retail trade 2.1 3.7 3.2 3.9 Transportation and communications 8.4 8.7 10.5 11.0 Financial services 15.0 18.9 24.9 29.0 Real estate, renting, and commercial activity 5.5 6.5 8.0 9.5 Education 2.6 2.6 3.5 4.6 Health 2.6 2.3 2.7 3.9 Social and other services 4.9 4.6 5.8 6.6 Source: NSS. For 2001, we used the above indicators to compare selected sectors and develop Figure 10.1. Another important element to consider in assessing competitive- ness is the exchange rate in comparison with Armenia's major trading partners (Table 10.5). Tables 10.1­10.5 illustrate the following:5 The number of new businesses registered per year declined between 2001 and 2004 in virtually all sectors (Table 10.1). At first glance this seems surprising, as the economy has been grow- ing rapidly and reform has been proceeding. However, further consideration reveals that it is plausible that business formation Strengthening Competition 287 Table 10.4 Gross Value Added Indicators, 2000­01 (% of the total value added) 2000 2001 Agriculture, forestry, and fishery 16.56 17.93 Industry (including mining) 32.67 31.50 Electricity, gas, and water supply 6.81 6.33 Construction 13.85 12.43 Wholesale and retail trade 13.00 14.71 Transportation and communications 9.45 9.39 Financial services 2.87 3.04 Real estate, renting, and commercial activity 0.40 0.45 Education 2.28 2.44 Health 0.43 0.38 Social and other services 1.68 1.39 Source: NSS. Figure 10.1 Business Competitiveness Indicators for Selected Sectors trade as % of GDP (75.00) 16.0 14.0 12.0 10.0 8.0 6.0 X X 4.0 2.0 0.0 average earnings employment electricity, gas, and water supply wholesale and retail trade transportation and communication financial services Source: World Bank Doing Business Indicators (http://www.doingbusiness.org). 288 Part II: Detailed Analysis Table 10.5 Foreign Exchange Rate Comparisons, 2001­04 (index) 2000 2001 2002 2003 2004 REER 108.7 105.5 96.2 86.8 90.8 REER-Russia 168.9 159.7 152.2 147.5 144 REER-United States 92.3 91.1 90.6 90.2 90 REER-Euro Area 116.2 117.1 118.6 118.8 117.6 REER-Iran 61.5 59.6 58.2 56.8 55.1 Source: IMF, International Financial Statistics (http://ifs.apdi.net/imf/). is slowing down. First, Armenia is emerging from many years during which the economy was centrally directed. There must be limited capacity among potential entrepreneurs to estimate realistic opportunities and to determine whether a market is satu- rated, what opportunities exist, and the extent of the competition. As experience is acquired, judgment will improve. Second, later sections of this chapter point out that the benefits of formality, in the sense of access to the financial system or accessibility of the legal system to contract and resolve disputes, is still very limited. Until reform improves these incentives, many entrepreneurs will prefer to operate informally. Entry into the wholesale and retail trade sector faces various constraints.6 High levels of employment but low average earn- ings characterize this sector. These features, coupled with high levels of GVA, indicate that the sector's anticompetitive practices are harmful to total welfare. Although high entry levels characterize the financial services sector--the entry-exit turnover and business formation indica- tors are high--its GVA figures illustrate that general activity is suppressed and the sector remains underdeveloped in terms of participating in the development of the Armenian economy. The high entry levels in the transportation and communications sector are explained by the large number of retail service pro- viders in the telecommunications sector and the recent increase in the number of lines in public transportation. Importantly, the procedure for granting ownership rights for specific lines remains anticompetitive. The increased number of different sector-specific services provided in the country also explains the high entry fig- ures (for example, freight). The electricity, gas, and water supply sector has moderate entry, GVA, and employment levels. Although there are issues still to be resolved, the sector's restructuring practice is one of the suc- cess stories in Armenia. Strengthening Competition 289 COMPETITION POLICY IN ARMENIA This section looks briefly at some of the issues related to sectors that are traditionally regulated in most economies, including telecommuni- cations, power, water, natural gas, transport, and airports. In Armenia, telecommunications regulation was brought under the jurisdiction of the Public Service Regulatory Commission (PSRC) in 2006 by legisla- tive change. The PSRC also regulates power, water, and natural gas. Other sectors are regulated differently. The General Department of Civil Aviation regulates air traffic. Yerevan Airport is operated under a concessional agreement by a foreign company. A concessional agree- ment was an important step in creating favorable initial conditions for successful air transportation regulation. The Ministry of Transport reg- ulates transportation. The railway company is still state owned, and unfavorable practices have been reported in this sector (see below). In addition, the Commission for the Protection of Economic Competition (CPEC) has authority regarding actions by companies in these sectors that are deemed harmful to the consumer. This authority has not yet been exercised, although the CPEC is adamant that it could exercise this authority in the future. The PSRC has been the recipient of substantial support from donors, both to provide technical assistance to upgrade skills and to improve its operating infrastructure. As a result, the capacity of the PSRC to regulate the industries under its jurisdiction has improved. Competition Law in Armenia A full analysis of the Competition Law is beyond the scope of this chapter. Furthermore, there are many aspects of competition in Arme- nia that, although relevant, are outside the ambit of the law. Nev- ertheless, some observations on certain features of the law and its application in Armenia are pertinent. Competition law is a relatively recent innovation in Armenia. The law establishing the State CPEC was passed in 2000 and the CPEC itself began prosecuting cases in 2001. Some key aspects of the law include the following: The law prohibits collusion among enterprises, although it fails to distinguish between vertical and horizontal collusion. Gener- ally, economists view horizontal collusion as potentially much more damaging to consumers. Vertical agreements can also present problems, but the circumstances under which this occurs need investigation and analysis. Potential problems could arise in Armenia because the CPEC's technical capacity for this analy- sis is limited. 290 Part II: Detailed Analysis Article 4 of the law prohibits anticompetitive practices between loosely connected groups, although the criteria for determining such behavior are loosely set out and would require intensive investigation, involving several different levels of complexity. Article 6 of the law indicates that a firm may be considered to have a dominant position in the market if it is not exposed to substantial competition, or if it commands a market share of more than one-third of total sales. The law forbids the abuse of dominant market power. In Arme- nia, the CPEC maintains a list of firms with such power and supposedly tracks them to ensure that their behavior does not involve abuse. Most of the cases brought by the CPEC have involved firms that are in this list. Cases have been prosecuted under Article 7 of the law, which applies to the abuse of a domi- nant market position. The Foreign Investment Advisory Service (FIAS) has criticized this approach on the grounds that the CPEC appears to consider that any firm in the list is automatically sus- pect--which essentially means any large company. Once a firm is on the list of dominant firms, scrutiny intensifies and transac- tions costs for the firm rise sharply. The CPEC has brought a number of cases of "unjustified pric- ing" based on a methodology involving comparisons between the selling prices of the company and general price indices. Such action makes little analytical sense, in that there is no allowance for shifts in relative prices, and it implicitly assumes that profits should not differ among products, companies, and sectors. Article 8 of the law defines market concentration and control through mergers and acquisitions and revolves around a market share of 35 percent or more. Combined with Article 10 of the law, which prohibits concentration unless it fosters competition, Article 8 is used to justify action against such companies. Article 9 requires that firms with substantial market power must be registered as being dominant. Implementation of the law presents a number of problems. The weak institutional framework in Armenia appears to result in some cases (especially in the areas outlined in later sections of this discussion) not being prosecuted, whereas other cases appear to show an excessive tar- geting of some firms. As in many other areas of private sector issues in Armenia, the law itself is not the primary problem. The existing leg- islation, while inadequate in some areas, could be used to prosecute uncompetitive behavior. Rather, problems lie in the application of the law, the protection that some groups appear to have, and the generally weak private sector business environment that limits opportunities for new entrants. These issues are explored at greater length below. Strengthening Competition 291 Telecommunications After independence, Armenia inherited a relatively extensive but low- quality telecommunications network. The design of the network was not commercially driven, which led to misallocated lines and low call volumes. These elements, combined with politically driven low local tariffs and in spite of high international call rates, resulted in low rev- enues per line. The outcome was outdated equipment, poor network quality, and a slow digitalization rate due to chronic underinvestment in the sector. During the early stages of restructuring, the government added substantially to the problems of the telecommunications sector by awarding a long-term monopoly (granted until 2013) to ArmenTel, the local telecommunications company, owned by the Hellenic Telecom- munications Organization (OTE), a Greek company. The process of privatization was poorly handled, was not transparent, and occurred without an adequate regulatory framework. In November 2004, the government issued a new license to Armen- Tel (after a longstanding dispute and renegotiations)7, signed a new Settlement Agreement with OTE/ArmenTel, and issued a second Glo- bal System for Mobile Communications (GSM) license with immediate effect, thereby partially liberalizing the market. The settlement agree- ment provides for a third GSM license to be issued on January 1, 2009. In addition, ArmenTel was deprived of its monopoly over Internet use in Armenia. At the same time, the government granted the company a monopoly for telephony based on the Voice over Internet Protocol (VoIP). The award of the second telecommunications license was not con- sistent with international good practice, given that only one candidate, K-Telecom (the operator in Karabakh using GSM technologies that is associated with Karabakh Telecom) was considered. The government perceived the issuance of this license as a "case of exceptional impor- tance" for Armenia's national security and provided Karabakh with international access to all telecommunications services. It is socially desirable to have an entrant in a competitive market for the following reasons: Entrants often offer a differentiated service not offered by incumbents. Entrants often provide existing services at a lower cost. Entrants often force the incumbent to produce services more efficiently. It is generally agreed (see Laffont and Tirole 2000) that an intelli- gent interconnection policy is the key to the harmonious development 292 Part II: Detailed Analysis of competition in the telecommunications industry. The offering of mobile services by ArmenTel and K-Telecom will give rise to two-way termination access charges. The government should therefore imple- ment interconnection charges and policies to facilitate competition, induce efficient network use, and encourage investment along with cost minimization. To ensure this end, the PSRC needs continued donor assistance to enhance its regulatory capabilities in the telecom- munications sector. Regulatory Challenge On the basis of the terms and conditions of both parties' licenses and the 2004 ArmenTel Settlement Agreement, it is unlikely that the cur- rent market configuration will change significantly and allow new entrants, at least not until 2009.8 In the early stages of partial market liberalization, however, embryonic competition if left to itself gen- erally favors the incumbent. Clearly, in the Armenian telecommuni- cations sector, competition is asymmetric: ArmenTel and K-Telecom compete for subscribers, but the former already has 205,000 mobile phone subscribers. Unfortunately, there have been few contributions (even theoretical) to the study of asymmetric competition, mainly owing to the compli- cations that arise by assuming that access charges should not be set reciprocally (see Cave, Majumar, and Vogelsgang 2002). The policy implications of the findings vary with the assumed underlying eco- nomic environment. Demand-Side Asymmetry and Nonreciprocal Termination Charges Demand-side asymmetry and nonreciprocal termination charges closely resemble Armenia's current situation in the telecommunica- tions sector. With the new law and its expectations regarding the regu- lation of the retail level, the implications may need to be modified; however, they are case sensitive and will require additional research. These implications include the following: Large differences in access charges (where the well-established firm is paid the higher access charge) act to reduce competition for subscribers. Two existing firms may differ over the setting of termination charges, and regulation is generally needed to resolve disputes. Generally, sector profits will increase with the increased difference in termination charges, but the gains to firms from this increase Strengthening Competition 293 change depending on the firm. This situation can lead to collusion between firms, with side payments from one firm to another. The market share of the networks affects total welfare. The mag- nitude of the effect depends on demand elasticity. In contrast to service providers, the increased access charges lower consumer welfare, which necessitates balanced access pricing policy in Armenia's telecommunications sector. The high cost of telecommunications in Armenia is particularly severe given that the country is isolated from Europe and the United States: there is an acute need for low communications costs in small, remote economies. It is critical, therefore, that the government and the PSRC deal with this situation appropriately and proactively. Utilities The electricity sector has been substantially restructured over the past few years. First of all, the distribution system has been privatized--an effort that appears to have been very successful. Whereas prior to privatization, inefficiencies, and large arrears strained the system to the breaking point, arrears have now been eliminated, electricity bills are collected on time, and the system is functioning reliably. This is one of the success stories of Armenian privatization. The power generation sector consists of five large generators, one of which is nuclear, two of which are thermal, and two of which are hydroelectric. One of the hydro generators and one of the thermal generators are run by Russia, which acquired the generators from the Armenian government in a debt-for-equity swap. Russians also run the nuclear station under a management contract. In addition to the large generators, there are a substantial number of small private hydro producers who supply the grid, which is compelled to purchase any power that they produce. The PSRC also regulates the natural gas distribution sector. Many of the same issues appear to apply to this sector and there is no indication that it was being run ineffectively. The Russian-Armenian joint-stock company "ArmRusGasProd" has invested intensively in developing a gas distribution network in the country. The CPEC regulates the producers9 on a rate-of-return basis. Although it is possible to quibble over the efficiency of rate-of-return regulation, a brief review has indicated that it appears to be carried out competently, although a more in-depth study would be needed to confirm this. The challenge will be to move economic regulation gen- erally onto more efficient, credible, multiyear, price-cap regulation. 294 Part II: Detailed Analysis Railways More than a decade of economic transition in Armenia is a lively illus- tration of forgone development opportunities owing to the transpor- tation blockade. Transportation is essential to promote sustained eco- nomic growth and prosperity, given companies' need to obtain inputs for production and deliver goods to the market. Armenia has suffered severely in this regard, not only because of the blockade by Turkey and Azerbaijan but also because productivity in the transport sector, particularly in railways, has been low. The evidence suggests that the operation of Armenia's state-owned railway is politically motivated at the expense of operational efficiency and social welfare. There have been reports of nontransparent and opportunistic price setting by the Armenian railway company. For example, during rate negotiations that the Ministry of Transportation and Communication was undertaking with the Gold Mining Company, the state-owned railway company attempted to extract very high profits through its dominant position in the sector by proposing high rates for transport- ing ore. The Gold Mining Company needed to connect its mining facilities (situated in the northeastern part of the country) with its melting facilities (located in the southwestern part of the country), and other means of transportation were economically inefficient. The min- ing company was in a weak position that the railway company was exploiting. Hence, the railway industry and the state-owned railway company in particular need to be restructured so as to contribute to sustained economic growth. The government of Armenia has decided to concession the railway and has selected the vertically integrated concession option. It will undertake the concession through a trans- parent and competitive international tender. A government decree dated September 16, 2006, appointed an Intersectorial Commission to present specific recommendations on implementation of the conces- sioning. A number of important issues related to the railway law, the concession design, and the oversight regime must be resolved to suc- cessfully concession the railway. The next section provides concepts to guide the process and ensure restructuring and regulation efficiency. Restructuring When considering rail transportation restructuring, Armenia needs to analyze the industry more generally within the context of its techno- logical and organizational features as well as evaluating the recent performance in detail. Five generic options for the vertical restruc- turing of railway are considered in the literature (see Kessides and Willig 1995). The competitive access option presented below has the Strengthening Competition 295 potential to bring more competition, market-based decision making, and innovation into the railway industry.10 Competitive access is char- acterized by the existence of an integrated operator11 required to make rail facilities, such as tracks and stations, available to other operators on a fair and equal basis through the trading of, for example, circula- tion rights. The potential for the implementation of the competitive access option in Armenia is favorable for a number of reasons. Active Intermodal Competition The passenger transportation data provided in Table 10.6 illustrate the small share of the railway sec- tor, which is mainly due to the enormous growth of motorization, to Armenia's landlocked geographic location coupled with its bad rela- tions with neighboring countries, and to the poor condition of the existing railway facilities (such as stations and passenger wagons). In freight, the expanding competitive trucking sector has gained the largest percentage of the transportation market. The figures on inter- nal freight volumes (see Table 10.7) illustrate the existing competitive pressures on the rail sector. This phenomenon is explained in terms of the following: Geographic size and road conditions inside the country. Virtually all of the destination points along the internal haul are connected by roads in good condition, which enables trucks to transport loads effectively and without delays. Industry dependence. Ore constitutes about 50 percent of deliv- ered (internal and imported) freight. Construction materials, both mineral and nonmineral (such as stones and cement), constitute Table 10.6 Percentage Share, by Weight, of Different Modes of Transportation, 1999­2003 Mode 1999 2000 2001 2002 2003 Passenger transportation Railway 1 0.8 0.8 0.8 0.6 Vehicle 73.2 75.7 80.3 82.7 85.5 Air transportation 0.4 0.5 0.5 0.5 0.5 Other (underground, trams) 25.4 23 18.4 16 13.4 Freight Railway 29.2 31.8 28 26.8 30.1 Vehicle 52.4 46.3 52.1 63.1 57.9 Air transportation 0.3 0.3 0.2 0.1 0.1 Pipeline 18.1 21.6 19.7 10 11.9 Source: NSS. 296 Part II: Detailed Analysis Table 10.7 Internal Freight in Volumes and Percentage Shares, 1999­2003 1999 2000 2001 2002 2003 Railroad Internal (thousands tons) 407.1 333.6 355.3 854.1 891.4 Internal (%) 0.29 0.23 0.25 0.42 0.42 Vehicle Internal (thousands tons) 2,157.6 1,776.0 2,203.1 4,323.4 3,548.4 Internal (%) 0.87 0.86 0.84 0.91 0.87 Source: NSS. another 15 percent. Therefore, an integrated operator would be unable to exercise its power on captive shippers (like the Gold Mining Company) if the competitive access restructuring is cou- pled with efficient regulation. In other words, Armenia's railway sector is far more contestable12 due to the strong competitive pressure from other modes of trans- portation. Economies of Scope, Coordinated Planning, and Reduced Transaction Costs It is often noted that the relationship among the supplied serv- ices, the rolling stock used, and the quality, quantity, and technical characteristics of the infrastructure is so close that these aspects need to be planned together. Thus, assigning different services to several operators may decrease the utilization of the sector's staff and physi- cal assets. This argument is especially relevant for the Armenia rail network, which is far from extensive. Increased Investment Incentives In a vertically separated structure, an infrastructure owner considering an investment in a facility with only one potential operator will anticipate bargaining away some of the benefit from the new service once it comes on line. The other problem that may arise with a vertically separated structure is the difficulty of the coordination of the operator with the infrastructure owner on investment decisions. A competitive access structure (with an integrated operator) is not dependent on such coordination, which reduces incentive problems. However, there is a possible shortcoming associated with the implementation of a competitive access option that needs to be considered: an integrated operator may have strong incentives to keep possible entrants out. Strengthening Competition 297 The solution to this problem is usually addressed by involving private management through concession contract arrangements and through regulatory policy (the latter being the subject of the next sub- section.) A private concession contract has the potential to stimulate the efficient development of the Armenian railway sector if the gov- ernment implements a competitive access restructuring. (Concessions are a broader form of lease in which the contractor agrees to make cer- tain fixed investments and maintains the use of the assets for a longer period.) A greater risk of closure of certain services, or of larger insta- bility, is obviously a potential issue with a private company. Again, concession systems reduce the risks inherent in the involvement of private enterprise. The design of a concession contract is a complex undertaking that should take into account, among other things, the type of contract (for example, horizontal versus vertical, passenger versus freight), award and duration (for example, type of auction, selection), content (for example, obligation, rights, asset ownership), price control, quality regulation, and so forth. It requires additional in- depth analysis of the sector, which is beyond the scope of the current report. But the quality of the concession contract will determine the design of regulatory policy and is therefore critical to the process. Efficient Regulation Although it is difficult to predict every situation that could undermine competition within the "competitive access/concession restructuring" option, Armenia should avoid two arrangements that affect incentives and deter efficient entry when designing its regulatory policy: Regulation should not constrain what an integrated operator can earn through cooperating with another entity. If earnings are constrained, the operator has a strong incentive to enlarge the portion of services it provides by hindering entry. Regulation should not permit firms to charge higher prices to shippers the more of their business they have, as this would also provide incentives for the firm to exclude other participants. These two situations suggest that Armenia should adopt pricing principles (widely discussed in the literature on regulation economics) that are efficient under the competitive access option. These pricing principles include the following: Ramsey pricing. This pricing scheme is appropriate with multi- service operation segments with actual or potential competitive pressure. The Ramsey approach apportions all unattributable fixed 298 Part II: Detailed Analysis and common costs of the railway among its services on the basis of their demand characteristics. The challenge for the regulatory body in implementing this policy is the complex task of estimating demand for the services provided by the integrated operator. Stand-alone costs. Where there is market dominance, Armenia should implement economically rational ceilings on rates, which can be obtained from the stand-alone costs. Stand-alone costs are the costs of serving any captive shipper if the shipper is isolated from the railway's other customers. The benefit of this pricing principle is that it overcomes the problem of entry deterrence possible under a competitive access restructuring option.13 Armenia should establish regulatory standards based on the principle that the integrated carrier that possesses a bottleneck should not refuse an agreement that provides full compensa- tion of all its costs, including opportunity costs. The standards should be implemented once disputes about predation through competitive access arise. Given that with railways there is a one-way access bottleneck, access fees should be priced using the principle of efficient com- ponent pricing. Pricing should be based on the idea that an inte- grated carrier should offer the services of its bottleneck at a price that yields it the same return as if it had performed the end user's services itself. Civil Aviation Issues relating to air traffic also provide grounds for concern. The national flag carrier, the state-owned Armenian Airlines, went bank- rupt and was liquidated in 2003. A local company, Armavia, bought the liquidated Armenian Airlines and acquired Armenian route rights for flights to and from Armenia via an agreement with the govern- ment. There have been reports that whenever foreign airlines operat- ing in Armenia have wanted to increase the number of their flights and have applied for permission to Civil Aviation, Civil Aviation has in turn sought the views of Armavia, which must be given greater weight by the Armavia agreement. This clearly substantially puts Armenia at a disadvantage. A small, remote economy that implements policies that drive up the price of communications (whether they are telecommunications policies or transportation policies) is effectively exacerbating the costs of its small size and remoteness. The Armenian national interest requires that the government take as bold a policy towards air transport as it appears to be taking with the telecommunication monopoly. An Open Skies policy should be implemented without delay, provided an arrange- ment can be made with Armavia.14 Strengthening Competition 299 LACK OF COMPETITION IN DISTRIBUTION Many private sector participants in Armenia operate in ways that would be unusual in the industrial countries. As is common in many countries with weak institutions, facilitation payments are frequently used as a means of ensuring that dealings with the public sector pro- ceed without undue delay. However, unlike other countries, in Arme- nia the system has to date not degenerated into one where massive corruption is pervasive, although undoubtedly there is corruption that extends to the highest levels. Rather, facilitation payments have evolved into a system in which interlocking obligations arising from favors and interventions govern much of the interaction among the business community as well as that between businesses and the legal system. A person who acts to intercede on behalf of another becomes a roof and the benefactor incurs an obligation to return the favor in one form or another at some point in the future. In Armenia, there appear to be few rules which cannot be modified or adjusted through the intervention of a roof, even within the judicial system. In sec- tors where a powerful roof exists, the incumbents appear to have the power to make life extremely difficult for new entrants. It is also noteworthy that foreign investors often do not have the network which provides them with a roof, and thus they frequently have more difficulty in navigating their way through the various bureaucracies and the judiciary. This reality has been reflected in all of the foreign investment surveys done on Armenia. In addition, companies that are only exporters claim that they are singled out for unfavorable treatment, especially with respect to VAT refunds. This phenomenon can act as a substantial barrier to entry in the economy, on the part of foreign investors as well as for new local businesses that attempt to compete with well-established companies and individuals operating in the Armenian economy. Although there have been improvements recently, subtle methods are used to impose severe operating constraints on new competition, ranging from prob- lems in clearing goods through customs, to inspections of various types by government officials, to outright sabotage of operations. Fur- thermore, such behavior is difficult to identify clearly because it is protected by all the members of a particular roof. There are numerous and widespread anecdotes to the effect that the economy is dominated by a few powerful groups that maintain a tight control on certain activities, particularly imports of petroleum, sugar, flowers, and wheat. It is, however, difficult to obtain hard evidence on the allegations, and the deeper the investigation is, the greater the contradictions are. There do appear to be roofs operating in these sectors but, on the face of it, there is some competition. On the other hand, there are also widespread allegations that companies in this 300 Part II: Detailed Analysis area are anonymously owned by interlocking share holdings, so that effectively the sectors are controlled by a small number of people who are making large profits and who engage in extremely conspicuous consumption. Petroleum Products Having no oil or natural gas resources, Armenia is totally dependent upon imports. Imports of fuel products constitute about 20 percent of total imports. This import group includes two major components in terms of volumes: petroleum products (including gasoline and diesel) and natural gas imports. Statistics on petroleum product imports appear to have been underreported. First, all of the data largely exclude imports by the Defense Ministry, Internal Affairs, and the nuclear power station, so it is not possible to determine the total imports of these prod- ucts. Furthermore, in spite of very rapid growth, statistics on gaso- line and diesel imports into Armenia show that in 2000 there was a sharp drop in import volumes and only a very modest growth afterwards. This occurred despite a very sharp increase in auto- mobile imports, which rough statistics indicate exceed 14,000 vehi- cles per year compared with 1,200 per year in 1999. Although oil prices rose in 2000, and demand may have been depressed by the fuel efficiency of recently imported automobiles and conversion of some motor vehicles from gasoline to natural gas, it does not seem likely that in the face of such rapid growth fuel consumption could go down. One explanation for the flat fuel imports is that there is smuggling of petroleum. The structure of market concentration underwent changes during 1999­2000. In 1999 there was only one gasoline importer in Armenia, so that market power was located inside the borders of the country. In 2000 and 2001 the number of importers into the country increased; looking only at the number of operators within Armenia, the mar- ket could not be defined as uncompetitive. However, analyzing the import chain reveals that one supplier provides the bulk (88 percent) of imported gasoline to importers. While the concentration of direct importers has declined, the virtually single sourcing of supply for gasoline importers means that the supply chain remains highly con- centrated. Even though it has been reported that one more company started to supply importers with gasoline in 2002, the share of the incumbent supplier firm remained dominant. It constituted about 70 percent of total supply to importers. (Noncompetitive behavior and its impact on the gasoline market and relationships to welfare losses are discussed in Box 10.1 and also shown in Table 10.8.) Strengthening Competition 301 Box 10.1 Noncompetitive Behavior in the Gasoline Market To stress the negative impact of noncompetitive behavior in the gaso- line market we made a very simple calculation of welfare losses (see text Table 10.8). (For the explanation of the terms used in the analysis below, see Tirole 1988.) We assumed a linear market-specific demand curve and estimated its slope and intercept considering price-increase versus demand-decrease scenarios during 1999­2000. Such a decrease in demand is mainly explained by increases in international oil prices in contrast to the observed nonprice impact (such as increased imports of cars after 2000) on demand during later years. It turns out that welfare loss, as a result of noncompetitive behavior in the gasoline market alone, was averaged at 1.3 percent of GDP. This loss figure can be even higher if we consider that there is a double marginalization (we have not included in the calculations the dead- weight loss, which can occur as a result of the monopoly power of one supplier of importers) and that the same situation dominates in many other similar markets (such as the diesel market). Moreover, the importance of liberalizing the petroleum oils market is accentuated by the fact that concentrated, rather high gasoline market profits (about 3.7 percent of GDP) have not contributed to reducing the unequal distribution of income in the country. Source: Authors' calculations. More interestingly, even though the large number of retail compa- nies operating in Armenia, ensures retail competition, the retail mark- up price was substantially higher than the imported (wholesale) price (which we took to be customs clearing price). The gross profit mark- up, most of which probably occurs at the wholesale level, constituted more than half of the retail price in 2001 (Figure 10.2). Since competi- tion at the retail level is strong, it is most unlikely that excess prof- its are being earned by retailers: most of the profits probably accrue to the wholesalers. These very rough calculations indicate that there appear to be large welfare losses arising from the concentration of the petroleum market, which could amount to the equivalent of over 1 percent of GDP. This loss figure would be even higher if we take into consideration the further profit margin earned by the single-source monopoly supplier of imports. Moreover, similar welfare losses can be expected to be experienced in other fuels, such as diesel. This is not included in the calculations of the deadweight loss, which can occur as a result of the monopoly power of the one supplier of imports. 302 Part II: Detailed Analysis Table 10.8 Welfare Calculations in the Petroleum Market, 1999­2001 1999 2000 2001 Retail price (drams per ton) 265,133.7 370,307.4 362,576.0 Customs clearing price (drams per ton) 118,300.0 145,100.0 159,200.0 Quantities consumed (tons) 258,100.0 181,400.0 187,500.0 Total profit (drams) 37,897,777,970.0 40,852,622,360.0 38,133,000,000.0 Estimated slope of the demand curve ­1.4 ­1.4 ­1.4 Estimated intercept of the demand curve 626,473.7 626,473.7 626,473.7 Quantities under customs clearing price 362,981.2 343,838.4 333,766.9 Deadweight loss (drams) 7,700,048,377.0 18,291,160,036.2 14,873,591,432.6 Deadweight loss (US$) 14,528,393.2 33,256,654.6 26,094,020.1 GDP (US$ million) 1,845.5 1,911.5 2,118.4 Deadweight loss as a percentage of GDP 0.8 1.8 1.3 Source: Author's calculations. Sugar The vast majority of sugar imports are brought into Armenia by a sin- gle importer. The sugar importing company was investigated by the CPEC, but before a judgment could be reached, the importing com- pany was dissolved and a new one was opened, but with the same shareholders. As a result, the original case did not proceed. Trade data indicate that 70 per cent of the imported sugar comes from the United Kingdom. Monopoly power in the sugar market is reflected in the price and markup data. Figure 10.3 shows that the difference between the cus- toms clearing price and the retail price of sugar is substantial--the retail margin is 47 percent, very high by any standards. As with petro- leum, there is substantial competition at the retail level so excess prof- its are unlikely to be earned at this stage. Table 10.9 reveals the patterns of the markup. Sugar has a very low demand price elasticity, which allows a high-profit-seeking importer with dominant power to hinder potential entrants. The sugar mar- ket profit was estimated to average around US$14 million during 1999­2002. Strengthening Competition 303 Figure 10.2 Gasoline Price Structure, 1999­2001 (drams per ton) 100 146,833 225,207 203,376 90 80 70 60 50 percent 40 118,300 159,200 145,100 30 20 10 0 1999 2000 2001 year margin over wholesale price customs clearing price Source: NSS. Figure 10.3 Sugar Price Decomposition (US$ per kilo) 100 0.21 0.22 0.18 0.18 90 80 70 60 0.25 0.26 50 0.24 0.25 percent 40 30 20 10 0 2002 2001 2000 1999 year markup customs clearing Source: Author's calculations. 304 Part II: Detailed Analysis Table 10.9 Sugar Market Data, 1999­2002 1999 2000 2001 2002 Markup (US$/kilogram) 0.185 0.176 0.222 0.205 Percentage annual change of markup -- ­4.650 25.933 ­7.349 Import of sugar (metric tons) 69,836 69,422 73,485 68,400 Percentage annual change in volumes of imported sugar -- ­0.592 5.853 ­6.921 Source: NSS. -- Not available. Some in Armenia argue that there are no large welfare issues aris- ing from the importation of sugar. They point out that the market is small and the main supplier can use its dominant position to purchase large quantities of sugar at a low price. Therefore, even if there are some monopoly rents being extracted, the welfare losses are offset by lower wholesale prices than would exist if there were several import- ers buying at higher prices in the exporting countries. The flaw in this argument is that if the current importer had concerns about new entrants, it would not exploit market dominance to the same degree and would extract lower rents. In addition, the large number of busi- nesses opening in areas where there is no restriction on competition illustrates the potential dynamism of the economy. Companies in the areas of ITC, restaurants, retail outlets, and building are being created in large numbers. Clearly, there is an entrepreneurial and dynamic element in the Armenian economy. Restricting competition harms this element as well as consumers. Wheat and Cut Flowers Gathering data in other markets proved more difficult, and it was not possible to undertake an analysis as detailed as that for petro- leum and sugar. Approximately 50 percent of Armenia's wheat is imported, as is the majority of cut flowers. There are widespread reports of the concentration of distribution in both of these prod- ucts. Potential competitors are being harassed at the borders by customs, which has employed many nontariff restrictions (such as health standards, quality standards, and valuation prices) in order to delay imports to the point where costs arising from dealing with the delay erode any potential profit. These issues are discussed at greater length in the next section. Strengthening Competition 305 HOW MONOPOLIES ARE MAINTAINED Corruption and rent sharing with state officials are a pervasive reality of doing business in Armenia. These practices affect the competitive environment in many ways. Market incumbents make illicit payments to or enter into collusion with governmental officials, agreeing to share rents, which are generated as a result of collusion. There are reports of state officials being given shares in the incumbent companies, which distorts the incentives associated with increasing public welfare through greater competition. The incentive for officials who are also shareholders in the incumbent companies is to maintain the status quo. As a result, the abuse of the state power is used as a tool for maintain- ing monopolies in the market. Although illicit and sometimes violent "private methods" have been used to maintain a dominant position in one specific market, the abuse of state power remains the most effec- tive and widespread method of maintaining monopoly power. The Customs Office Armenia has made great strides in attaining compliance associated with WTO access. This has not prevented the Customs Office from remaining one of the predominant instruments for hampering or elim- inating potential competitors, particularly in the distribution sector. Customs clearance and valuation procedures are far from being trans- parent and standardized (see Box 10.2), with substantial discretion being applied in the licensing of brokers, storage, freight transporters, and the operation of the free trade zone. In addition, the continued widespread use of reference pricing in contravention of WTO commitments is another method of delay- ing imports. Since customs also has some responsibility for enforc- ing product standards (for example, on health and safety grounds), the number of potential administrative barriers is legion, particularly since some of these standards are also under the jurisdiction of the Ministry of Health. There are therefore numerous opportunities for delaying imports for a number of reasons. Some of these issues are discussed in greater detail in the following sections. Limiting the Number of Customs Brokers and Associated Activities The Customs Office controls "unwanted" private enrollment in cus- toms operation through the widely reported unfair and discretionary licensing of customs brokers.15 The Customs Office appears to strictly limit the number of brokers through customs examinations. Each year 306 Part II: Detailed Analysis Box 10.2 Anti-Competitive Practices of the Customs Office Royal Armenian JV LTD imports 8,000­9,000 tons of green coffee annu- ally, processes it, and exports 70 percent of the processed product (mostly to Europe). The company was founded in 1998, and 96 percent of its capital belongs to a private person, a citizen of the Republic of Cuba. On July 2004 the chairman of the company publicly accused the State Customs Committee Deputy Chairman and other senior officials of personally soliciting a bribe last year in return for undervaluing the price of imported coffee beans, thereby reducing the import tax the company would be required to pay. After the company refused the offer, customs retaliated by overstating the value of the coffee imported at US$1.8 per kilogram; the actual cost was US$1.2, and the estimated cost for other importers was calculated to be less than US$1 per kilogram. Royal Armenia won a legal action against the Customs Committee for discrimination, but the ruling that the increased tariff on their imports should be reduced has not been enforced. The chairman of the com- pany was later reported as having retracted his accusation regarding the customs authorities. Source: Based on interviews and newspaper reports. there are examinations for customs brokers: even incumbents must take annual exams.16 A reliable source reported to the mission that in 2003 apparently 30 out of 36 applicants passed the examination but after internal review the Customs Committee declared that there were only 5 successful applicants. One of the unsuccessful applicants (an operating broker) was apparently able to get another minister to call the chairman of the Customs Committee on his behalf, so that finally there were 6 successful applicants. On the face of it, the procedure is entirely valid--a written examination subject to scrutiny and appeal. In practice, however, the results of the examination are subject to internal review and adjustment at the whim of the Customs Committee. This "informal review" procedure is not written down anywhere. Since there is such wide discretion, however, the Customs Committee can strictly con- trol the number of brokers and can ensure that they are compliant through the annual examination. Clearing Procedures Customs have introduced electronic clearing procedures to expedite the clearance of goods. Customs brokers, however, do not have access Strengthening Competition 307 to the computer system and therefore have to deal with paper doc- umentation that nullifies many of the advantages of computerized clearance and that also keeps the brokers in the dark with respect to the status of the goods that they are attempting to clear. The result is that there are uncertainties in the process and delays that can cost importers substantial amounts, especially since they are charged for the storage of the goods that are held up. In late 2005 the customs authorities began piloting a direct trader input system and reforms towards automation in clearances in goods covering about 5 percent of imports. They undertook to extend the pilot to about 70 percent of imports by April 2006. If they are implemented successfully and in line with international good practice, these steps would constitute a major reform in customs administration and significantly benefit the forces of competition. Another issue frequently cited by importers is the continued use of reference pricing by customs unless extensive documentation accom- panies shipments.17 If any of the documents are missing, then customs applies reference prices to the shipment, a practice that is not consist- ent with WTO rules. In other countries goods can be cleared with partial documentation under surety bonds that provide a guarantee in the event that the declaration was incorrect. However, the lack of financial system development in Armenia implies that surety bonds are difficult to obtain and that even when they are obtained, the cost is high. The problem of reference prices is further compounded by the relatively high VAT rate (20 percent) and the historic difficulty in obtaining drawbacks of VAT payments upon the export of any goods that use imports as inputs. However, a significant improvement has taken place in the practice of refunding VAT dues to exporters, in the course of 2005, both as to amounts and as to timeliness; this trend should continue. The Role of the CPEC Should the CPEC be one of the pillars of the policy of promoting com- petition in Armenia? There are understandable concerns that it could become yet another layer of bureaucracy that leads to "inspections" and interference in the ability of businesses to function, which in turn could be used to harass entrepreneurs and to protect incumbency. Arguments in favor of the strengthening of the CPEC are based on the recognition that it could be one of the ways in which the restrictions on competition that abound in several sectors of the economy could be resisted. The view of this chapter is that, in its present form, the CPEC is not a significant bulwark against anticompetitive behavior. It has neither the necessary skills, nor the staff, nor the facilities to operate effectively. Although it has been successful in a limited number of cases,18 it is not 308 Part II: Detailed Analysis a force in areas where competition is obviously restricted. Its report on the petroleum market is deficient in depth as well as analysis.19 This does not mean that attempts to assist the CPEC therefore should be abandoned, since over the longer term the commission could be one of the instruments for promoting competition in Armenia. Modest technical assistance is warranted and perhaps some help with improv- ing its ITC resources. It is unrealistic, however, to perceive the CPEC as useful in the shorter term. It needs several years to evolve into an instrument against restrictions on competition. Nevertheless, its con- tinued existence is assured, and the interests of the private sector lie in its improving its technical competence. BARRIERS TO ENTRY Formal barriers to entry have been reduced. For example, the cost of registering a business is now much reduced and occurs with few delays. The Armenian authorities are determined to eliminate delays altogether. However, reducing the costs of formalization does not address the many issues related to the benefits of being formal. As the next sections will point out, public goods related to the business sector are still inadequate. To enhance competition, the authorities are focusing on access to finance and access to the system of for- mal contracting. In Armenia, financial markets function poorly and the system of contracting and dispute resolution is haphazard at best and under normal circumstances is not effective. It does not provide Armenian businesses with incentives to formalize, particularly since it also allows petty officialdom to target formal businesses with inspec- tions that have payoffs as their aim. There has been some progress in reducing these inspections. For example, tax officials can only conduct one audit per year. However, they are still allowed to undertake "fact finding," which is not clas- sified as an audit. Nevertheless, businesses report that information obtained during fact finding, which can occur as often as requested, is used in annual audits. The protection of the restriction on auditing is therefore no more than limited. Financial Markets The relationship between financial market development and pov- erty alleviation is well established (see, for example, Holden and Prokopenko 2001). Financial markets play a key role in intermediat- ing between savers and investors. In addition, the financing of new projects enhances openness and competition by ensuring that prof- Strengthening Competition 309 itable investment opportunities do not go untapped. Although it is currently difficult to measure the overall severity of anticompetitive practices owing to inefficient financial markets, two practices do put pressure on Armenia's competitive environment. First, in product markets where a firm (usually with power or con- nections) has propitious access to or controls over financial instruments, the incentive is created to use them to affect the product market and thus hinder possible entrants. Second, there is indirect or financially driven deterrence. Because of the vulnerability of the financial sector Armenia's lending institutions are constrained by two lending options: either to be particularly cautious in entering into financial contracts with an incumbent firm or to restrict lending to other business entities in the same or a closely related product market as their returns become more or less sensitive to the effect of product market competition.20 In both cases there is excessive pressure on product market competition. Financially driven entry deterrence is most important when financial market competition is limited. This situation establishes an important policy implication: it is difficult to promote product market competi- tion if financial markets are concentrated? This in turn suggests the need for more antitrust coordination among the regulators of financial and product markets than is currently the norm in Armenia. Financial Underdevelopment Financial markets in Armenia are bank dominated, underdeveloped, and effectively finance neither local production nor foreign trade. Figures 10.4 and 10.5 show the level of financial market devel- opment in Armenia compared to that in other countries at various stages of development.21 Even by the standards of low-income coun- tries (those with per capita incomes of less than US$1,000) the ratio of credit to the private sector is very low--only one-fourth that of the average credit-to-GDP ratio in other low-income countries. The lack of finance for business acts as a severe barrier to entry. Potential competitors cannot obtain finance for their businesses, either in the form of working capital or to finance foreign trade. The result is that incumbents' market dominance is rarely threatened by new entrants and that wealth remains concentrated among those with resources. In spite of the low amount of credit to the private sector, there does not appear to be an excess demand for loans. On the contrary, at cur- rent rates of interest there is an excess supply of loanable funds. For example, the largest bank in Armenia, Hong Kong and Shanghai Bank- ing Corporation (HSBC), lends a much smaller proportion of its liabili- ties (10 percent) to private business than it does in other countries in which it operates, where it lends 60 to 70 percent of liabilities. 310 Part II: Detailed Analysis Figure 10.4 Interest Rate Spreads (lending rate minus deposit rate) 60 Armenia 50 low income lower middle income upper middle income 40 high income countries cent 30 per 20 10 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 year Source: IMF, International Financial Statistics. http://ifs.apdi.net/imf/. Although interest rate spreads between borrowing and lending rates have fallen, they remain higher than in other low-income countries. Financial activity in the foreign trade sector is limited. Some import- ers do use letters of credit in the banking system, but the practice is not widespread. There is little export financing. The practice of using irrevocable letters of credit as a basis for providing funds to fulfill export orders appears to be nonexistent. As a result, the financial sys- tem cannot be said to support the growth of trade, which in turn is the key to promoting more competition in Armenia. The essential skills in raising letters of credit are confined to a small number of banks. Before exporting under letters of credit can occur, training in the minutiae of the procedures will be required for most banks in Armenia. Financial sector development is constrained by weaknesses in financial infrastructure, such as creditor rights, credit information, and corporate governance. Weak Creditor Rights Weak creditor rights raise losses in the event of default, contributing to wider lending margins and reducing the availability of credit. This Strengthening Competition 311 Figure 10.5 Domestic Credit to the Private Sector (percent of GDP) 160 140 120 100 80 percent 60 40 20 0 1997 1998 1999 2000 2001 2002 2003 year Armenia high income upper middle income lower middle income lower income Source: IMF, International Financial Statistics. http://ifs.apdi.net/imf/. weakness arises from deficiencies in the law and practices govern- ing collateral. To enhance creditor rights, the authorities will need to address the following issues. Secured Transactions Framework. In Armenia, movable property cannot be used effectively as collateral to secure loans.22 There are numerous problems with the secured transactions framework. Float- ing pledges are not allowed and there are no registries that allow pledges to be perfected. Repossession is time-consuming and costly. A particular problem with financing exports is that there is no provi- sion that allows the use of future production as collateral and thus financing against export orders is not feasible. A similar problem arises with imports because there is no provision for pledging goods that are not in the possession of the borrower: goods that are being imported, even though they have been paid for by letter of credit, 312 Part II: Detailed Analysis cannot be seized in the event of default. The effect of the inadequacies of the collateral framework extends throughout the economy, with the result that banks correctly perceive lending as extremely risky.23 To understand the necessary conditions for establishing a framework for secured transactions, it is essential to understand the four economi- cally important stages of any system, which use the following: 1. Creation. The process by which the creditor establishes a security interest in a specific property (the collateral) 2. Priority. The process by which the lender establishes the priority of the security interest 3. Publicity. The process that makes public the priority of the secu- rity interest 4. Enforcement. The process by which, upon the debtor's default, the creditor will seize and sell the collateral to satisfy its claim Each of these stages must function effectively for collateral-based lending to occur. The authorities are aware that currently none of these stages work well in Armenia and that they will need to ensure that reform supported by the World Bank as part of the PSRC, USAID, and other international financial institutions on collateral framework is implemented effectively. The government has adopted a plan to strengthen creditor rights, improve the civil procedures for debt recov- ery, and simplify the procedures for the certification for property sales and the registration of secured credit. The government has established a working group to review six draft laws in this connection which would establish the required legal environment for strengthening creditor rights. Commercial Law and Contracting The law governing commercial transactions is inadequate, with the result that contracts tend to be informal and to take place between people who know each other well. The effect is a barrier to entry-- arm's length contracting is a requirement of a well-functioning eco- nomic system. It also has the effect of increasing incentives for infor- mal behavior, which is widely observed in Armenia. Problems include the following: Decisions on issues related to commercial transactions are based on an agglomeration of civil code laws, which makes the appli- cability of contract provisions uncertain. Contractual agreements are based on "unless otherwise governed by law" rather than on "unless otherwise stated in the contract" Strengthening Competition 313 provisions. Since many commercial transactions are subject to a wide range of laws rather than a more unified commercial code, this provision requires full knowledge of all the laws and imple- menting regulations that may apply to an agreement, which com- pounds the problem of the unavailability of translations. Not only does precedent not play a role in court decisions, but there is no record of cases being kept, although this is partly being addressed by a World Bank legal reform project. On top of the uncertainty arising from the commercial code, judges are widely viewed as incompetent and corrupt. As a result, businesses have little recourse if they are the victims of anticompetitive behavior. Notaries and Formal Contracts The role of the notaries adds to the barriers to entry and the incentives for informal behavior. Notaries (appointed by the Minister of Justice) provide public services and are required to validate most contractual documents. Procedures are antiquated and slow. Notaries also act as judges and determine the validity of documents (and in many cases insist on preparing the documents). Notary fees and stamp fees are high and are based on the value of the transactions covered by the con- tract being notarized. The result is that many contracts remain infor- mal because of the high cost of notarization. This situation also favors incumbents who have experience with working with one another. Credit Information The credit registry, which was introduced by the Central Bank in January 2003, aims to reduce credit risk by the crea- tion of an information system on the creditworthiness of customers of banks and credit organizations operating in Armenia. However, while the credit registry has been beneficial to banks' supervisors, this information is limited to debtors of banks only. A private Credit Bureau began operations in early 2004 and has focused in its first year on obtaining information on debtors from the banks, from other finan- cial institutions, from utility companies, and from government offices. This information should improve the ability of creditors to evaluate prospective borrowers' creditworthiness. However, the Credit Bureau has faced difficulties, because most banks are reluctant to provide information for free and to then have to pay for services. Furthermore, based on tradition and culture, Armenian borrowers are reluctant to share their information with others, and ensuring privacy protection substantially constrains the effectiveness of the Credit Bureau. It is currently revisiting its operating model and expanding ownership to include banks, with a view to remedying these obstacles. 314 Part II: Detailed Analysis In countries where credit bureaus have been established success- fully, the central bank plays an important role in creating incentive mechanisms for commercial banks to share information on their cus- tomers. The analysis suggests several situations in which Armenia's Central Bank could play an active role in facilitating private credit bureau development: Credit bureau reports should be made a requirement for lending coupled with mechanisms to ensure efficient entry and exit by bureaus to avoid abuse of their exclusive rights. A borrower's explicit consent should be required prior to any- one's gaining access to their file, to ensure better privacy protec- tion. Other actions that may prevent privacy violations include restrictions on banks for gathering certain kinds of information, the customer's right to access account information, validation of one's own files, and elimination of individual files after a certain period. Access to information by borrowers and credit institutions should be granted on the basis of membership. In other words, the prin- ciple of reciprocity should guide credit bureaus and should be stated in the contractual agreement between the bureau and credit institutions. Membership should not be based on fees in any form. However, credit bureau activity should be profit-ori- ented. In cases where a member provides inaccurate information or fails to provide data, sanctions (ranging from fines to loss of membership and hence denial to the bureau's files) should be imposed. Credit bureaus are exposed to potential conflicts of interest, especially if they are owned by a group of lenders (each lender wants to exploit the information provided by other lenders with- out disclosing his/her own). Therefore, Armenia should adopt a balanced-ownership policy. For example, credit bureaus could be incorporated as private companies and owned by a consor- tium of lenders to create the incentive for information exchange. Alternatively, independent ownership, coupled with a proper membership policy, could serve as an incentive to lenders to exchange information. Apart from the efficiency debate and the coexistence of different institutional arrangements, there is concern regarding the opti- mal amount of information sharing and its content. In contrast to relying on the general statement that information dissemi- nation reduces adverse selection problems (due to bad risks in the population of credit seekers) and makes the information on which banks base their lending decision homogenous, the policy Strengthening Competition 315 for the development of an efficient credit-information sharing system should be focused on the regulation of the quality of the disclosed information. Furthermore, the authorities should work to enhance corporate governance, which is weak. While company law is sound, imple- mentation is weak. Many corporations and some banks are not listed, and they may have relatively few owners who do not see the benefits of adopting the governance structures covered by the law (that is, open and closed joint-stock companies) that are intended to ensure good corporate governance. There is a need to broaden the coverage and enhance the enforcement of company law to strengthen disclosure, accounting, and oversight by corporate boards of directors. Corporate Governance Corporate governance in Armenia is weak owing in large part to the lack of transparency of ownership and con- trol in the banking and corporate sectors. The Detailed Assessment of Compliance with the OECD Principles of Corporate Governance indi- cated that while company law governing open and closed joint-stock companies is sound, implementation is weak. Only a limited number of corporations and banks choose to operate under this law, which is covered under securities legislation and enforced by exchanges and securities commissions. Many entities are closely held, are not pub- licly traded, and may have few owners who do not see the benefits of adopting a more formal governance structure. Hence, the enforcement of sound corporate governance practices through securities legislation and securities regulators is weak. There may be a need for a fundamen- tal reform to enhance the coverage and enforcement of company law if corporate governance, disclosure, and accounting are to improve. The opaque ownership structure of the corporate sector makes it more difficult for the banking sector to assess the risks on corporate sector exposures. Official disclosures of direct and indirect owner- ship are substantially weaker than in other transition economies. No disclosure of beneficial owners is required under Armenian law. This limits the ability of banks to ensure compliance with prudent limits on loan portfolio concentrations and related party transac- tions. Although legislation requires all companies to prepare their financial accounts in line with the Accounting Standards of the Republic of Armenia, in practice these standards are not observed or implemented, partly because of lack of training. Steps should be taken to improve accounting and auditing practices and finan- cial reporting requirements and to strengthen public information, including information regarding beneficial owners of banks and publicly traded companies. Furthermore, the supervising boards of 316 Part II: Detailed Analysis Armenian companies are not sufficiently effective or accountable. The company law should establish the separation of company own- ers from company managers and should strengthen and clarify the fiduciary duties and accountability of boards of directors. These shortcomings in corporate governance can be addressed through legal, institutional, and supervisory reforms. Given the weak implementation of company law, the Central Bank has developed a proposal to accelerate improvements in corporate governance in the banking sector through legislative and supervisory means and by introducing upward-consolidated supervision to address exposures to beneficial owners. There are a number of specific reforms applying to the corporate sector that should be implemented to address the shortcomings in corporate governance. They include improvements in accounting and auditing practices; financial reporting requirements; enhanced disclosure of the shareholders of corporations; improved access to information by the public, in particular by making the com- pany registry publicly available; and strengthening the role of boards of directors. Bureaucracy A further problem for businesses in Armenia is the failure of the gov- ernment to issue VAT refunds to those companies that are primarily or solely engaged in exporting. For example, a company that exports smelted copper is owed large refund amounts that stretch over more than 12 months. While this is not an issue that can be said to impinge directly on competition in Armenia, it is another illustration of the failure of the state to protect property rights and the rights of busi- nesses, and it is a further example of the limited benefits of being formal. Similar problems are reported by one of the major hotels in Yerevan, which is having difficulty getting VAT refunds on the cost of building. A recent FIAS Report (World Bank 2004b) highlights the role of the tax authorities in raising transactions costs. Over 80 percent of respondents indicated dissatisfaction with their dealings on tax issues. Within the tax system and tax administration the most problematic issues are the "extra-legal requirement for advance payment of taxes," the "frequency of changes in rules and rates," and the "availability of information regarding the laws and regulations." While these data must be interpreted cautiously (Armenian businesses are not known for their compliance with the tax codes), the widespread dissatisfac- tion with the lack of transparency of the process is indicative of the extent to which taxation can be used as an instrument for harassing entrants into areas that are dominated by incumbents. Strengthening Competition 317 CONCLUSIONS AND POLICY IMPLICATIONS There is little doubt that there is a serious lack of competition in some sectors of the Armenian economy. The preceding analysis indicates that in several sectors, particularly those related to imports and dis- tribution, monopoly profits are being earned by a small group that is profiting greatly and imposing welfare losses. This chapter concludes that the concerns regarding lack of competition are warranted. It iden- tifies three separate areas of the economy where competition, or the lack of it, is an issue. These areas are the following: the traditional sectors that are regulated--utilities and natural monopolies areas in which competition is widely perceived to be restricted, especially petroleum, sugar, wheat, and flowers barriers to entry that maintain the position of market incum- bents Nevertheless, it is important to recall that Armenia has been among the fastest-growing economies in the world for several years. This rapid growth is all the more noteworthy because it has taken place against a backdrop of closed borders. Lack of competition, however, carries long-run efficiency implica- tions as well as the danger that the apparent concentration of wealth could reduce political support for reform and limit the benefits of rapid growth. There is ample evidence of poverty in Armenia, even in Yerevan, and outside of the capital it is far worse. Poverty reduc- tion requires that the benefits of growth are realized by the mass of the people. The more that competition is restricted and activities are reserved for incumbents, the less likely it is that the benefits of growth will be widespread. There are additional unfortunate consequences of the dominance of the oligarchs in some sectors. The concentration of wealth pro- vides substantial ability to grant "favors" to officialdom to subtly (and sometimes not so subtly) hinder competitors entering the market. The phenomenon of roofs, discussed in the section on business practices, is one manifestation of the way in which incumbents maintain their market power. In other cases, it appears that there are direct attempts to prevent competition. Widespread subversion of the "rules of the game" by the rich and powerful sends strong messages to those who are on the outside. It contributes to a general attitude that taxes should not be paid, officials should be bribed, and goods should be smug- gled. This does nothing to develop the foundation for a modern com- petitive economy. 318 Part II: Detailed Analysis Geography and Competition Policy The Armenian economy is small and remote. One of the aims of policy should therefore be to ensure that nothing exacerbates these geographic facts of life. In particular, policy should ensure that com- munications and transport are as low cost as possible. In the past, this principle has not been adhered to and almost appears to have been designed to intensify the disadvantages of Armenia's geographic limi- tations. The ill-conceived awarding of the telecommunication monop- oly to ArmenTel has substantially raised the cost of communications. The policy regarding regulating flights into and out of Yerevan raises the cost of traveling to and from the country. There is no valid eco- nomic reason for these restrictions. The government has started to take action regarding the telecommunications issues. It should take similar measures with regard to air transport and should declare an Open Skies policy, allowing any airline that wishes to fly to Armenia to land at Yerevan. Regulated Sectors In the sectors that are traditionally the object of regulation, progress is being made. Actions are being taken to deal with the ArmenTel monopoly, and telecommunications regulation will come under the jurisdiction of the PSRC once pending legislation is passed by Par- liament. Regulation of power generation and distribution, as well as natural gas, has advanced. Both of these industries are also regu- lated by the PSRC. The functioning of the PSRC appears to have improved--for example, they have reversed price increases in the energy sector. Although it is possible to quibble about the methods used as well as some of the details of regulation, there is no doubt that in this area significant progress has been made. There are concerns regarding regulation in sectors that are still state-dominated, especially the railways, which are a vitally impor- tant industry given Armenia's transport problems. Currently, railways are both administered and regulated by the Ministry of Transport and Communications. Transport-intensive industries such as mining complain that attempts are being made to raise tariffs without any apparent justification. No options should be dismissed without con- sideration, including privatization of the railways and bringing regu- lation under the purview of the PSRC. Regulatory capabilities are scarce in Armenia, and consolidation under one body, which could then be the recipient of intensive technical assistance, could be the best policy option. Strengthening Competition 319 Areas Where Competition Is Restricted This chapter examines sectors of the economy that are widely regarded as suffering from lack of competition through monopoly and the restriction of competition in the distribution chain. It concludes that there is evidence that prices are higher than they would be if there were more competition, and that welfare losses are significant. Most of the sectors discussed are involved with the distribution of imported products. In one sense, the problem would disappear immediately if there were the ability to import freely. However, the chapter points out that customs has become an instrument for maintaining the monop- oly position of incumbents. There are several things that can be done about this situation. The first is to increase the transparency of the process. When light falls upon irregular practices, they tend to decline. Two steps are pos- sible in the short run. First, make the computerized customs data- base available to brokers, including the clearance of all goods, so that restrictions on shipments must be explained. Second, make the cus- toms examinations more transparent by following the letter of the law in this regard. All examination papers and the marking of the papers should be available for inspection. Whether this is feasible from a political economy perspective is another matter. Recommendations for the reform of customs have been made by donors almost from the time of the country's independ- ence. Some progress has been made, particularly with respect to the procedures related to Armenia's WTO accession. Nevertheless, there is still a long way to go in this regard and the process is clearly the main way in which monopoly positions are maintained. Since the number of access points to Armenia is limited as a result of closed borders with Azerbaijan and Turkey, the effectiveness of restricting certain imports is enhanced. In addition, the use of reference pricing is still widespread, in contravention of WTO undertakings. Although cus- toms and revenue officials claim that it is necessary to check invoices in order to ensure that there is no cheating on VAT payments, there are alternative procedures that allow goods to be cleared pending a check on prices that are used successfully in other countries. There is no reason why these methods could not be adopted in Armenia. Another avenue is open to the authorities to improve competition in the petroleum market. They could require that a high proportion (perhaps 50 percent) of the petroleum imported into Armenia must come from one of the large Western petroleum companies and that the invoices supplied with the shipment are made available for public inspection. Such a step would introduce contestability into the market and would improve transparency. 320 Part II: Detailed Analysis The CPEC In its present form, the CPEC's ability to ensure competitive markets in Armenia is limited. It lacks the resources and the skills to be effec- tive, and it has no power to enforce its own judgments except through the court system, which itself is arbitrary and open to outside influ- ences. The policy question is therefore whether devoting resources to strengthening the CPEC is warranted. Such assistance would take the form of upgrading the CPEC's resources, particularly in the ITC area, upgrading the skills of the CPEC, and perhaps giving it the power to enforce its judgments. Concerns regarding strengthening the CPEC rest on the under- standable desire to avoid yet another layer of bureaucracy that could be used as an instrument to harass business. However, the CPEC is not going to disappear; its existence is enshrined in law and it appears to have powers that so far have been untested. Upgrading its skills is necessary to prevent it from turning into the very institution that some fear it could become. The more professional the CPEC is, the less likely it will turn into an instrument for restricting rather than promoting competition. The Long-Term Promotion of Competition and Development In the long run, promoting competition in Armenia is inseparable from promoting private sector development in the country. The ability to restrict competition frequently arises because the institutions that underlie private sector activity are underdeveloped, providing strong incentives for informal behavior. In Armenia, the public goods that provide the foundation for private sector development are weak. The court system does not function effectively, and arm's length contract- ing is risky, so that transactions tend to take place among those who know and trust each other. This gives powerful advantages to incum- bents. Similarly, the financial system remains underdeveloped even by the standards of low-income countries. Hence, those who have substantial financial resources are in an especially strong position to maintain and strengthen their market dominance. The development of institutions supporting the private sector is the only long-run solu- tion to promoting a competitive market environment. Legal reform, the strengthening of property rights, and corporate governance are some of the measures that will greatly enhance private sector activity. In addition, the institutions of government must be upgraded. They need to be of positive assistance to the private sector, rather than being the instruments for maintaining anticompetitive behavior that they are now. In particular, the reform of the customs service is the key to promoting competition. Strengthening Competition 321 Shorter-Term Measures Several of the suggestions in the preceding sections could be imple- mented with little delay. The announcement of an Open Skies policy would signal the government's commitment to establishing a more competitive business environment. An immediate change in the exam- inations for customs brokers would signal a commitment to transpar- ency in customs procedures, as would giving brokers access to the computerized clearing system. Another measure that would promote competition would be to insist that a certain percentage of imports of petroleum be purchased from one of the large international petroleum companies and that the documentation be made publicly available. Suggestions for Further Work There are a number of areas in which further work is needed in order to suggest more detailed reforms. The most important of these are the following: A thorough review of the commercial code is required in order to identify areas that are inimical to modern commercial prac- tice and to suggest reforms. This review should be undertaken not only by lawyers but also by a team consisting of a lawyer and an economist in order to ensure that the incentives inherent in the code are identified from an economic as well as a legal perspective. A similar review of the consumer protection law is warranted in order to take stock of the system for consumer protection that underlies the role of the CPEC. Such a review would ensure that the CPEC does not become an instrument for the harassment of businesses and that the incentives in the law make sense from an economic perspective. This review could also look into the feasibility and desirability of establishing a small claims court where consumers and producers who have been harmed by anti- competitive behavior could sue the perpetrators. At this point, a stocktaking and review of the regulatory frame- work is also warranted, to determine how well it is now func- tioning, where further strengthening is needed, and how the PSRC can be further supported in its work. A thorough review is needed of the issues related to secured trans- actions reform, from an economic as well as a legal perspective. This should include the feasibility of using a similar approach to that taken in Romania (see Annex 10.1 to this chapter). ANNEX 10.1 An Example of Successful Secured Transactions Reform in Romania R ecent reform of the secured transactions framework in Romania has transformed the lending environment by facilitating the use of collateral as security for lending, not only by the banking system but also by equipment suppliers, wholesalers, and agricultural sup- pliers. There were many similarities between the Romanian financial sector before this reform and that in Armenia: severe financial under- development, the inability of a large sector of the economy to access credit, and a distrust of banks. Therefore, use of this model could have significant potential for the development of the Armenian financial system in a way that could substantially reduce barriers to entry. The essential elements of a well-functioning system of collateral provide for the creation and registration of collateral as security for a loan and its rapid repossession in the event of default. The Roma- nian reform involved first changing the law and legal institutions and second establishing a privately run registry where security interests regarding the pledging of property as collateral for loans could be recorded. This was done within the context of the civil code system that exists in Romania, which has further lessons for Armenia. The main features of the reform included the implementation by the government (with the support of the World Bank) of a Law of Secured Transactions that permitted movable property, both tangi- ble and intangible, to serve as collateral for a loan. The second step involved setting up the filing archive to permit the law to operate by recording pledges of property and establishing priority regarding which creditors have the first rights to repossess and sell the collateral in the event of default. 322 Strengthening Competition 323 There are a number of innovative features of the secured transac- tions reform: The law abrogated all existing legislation affecting debt, so that there was no danger of ambiguity regarding the validity of the pledges. The filing archive in which pledges are recorded is run by an association of lenders, so that the public sector is not involved. The archive is electronic and priority is determined electronically at the time when pledges are recorded. No documents are necessary to file a security interest. This allows Internet-based filing, currently the only filing archive in the world to have this feature. The effect is to broaden coverage, particularly in the rural areas, as well as to reduce costs. No notaries are involved in the process. Repossession takes place outside the court system. If repossession is not disputed, the creditor can simply collect the pledged prop- erty. Lenders confirmed that repossession was not a problem. In the event of dispute, upon evidence of the validity of the debt, which does not require proof beyond the entry in the filing archive, an officer of the court can seize the pledged property without the necessity of a court hearing. Very harsh penalties for wrongful repossession discourage credi- tors from abusing the system. The new law and filing archive are among the most modern in the world, even among the developed countries. Prior to the new law, credit in Romania was scarce and expen- sive. The Romanian financial sector was underdeveloped even by the standards of many transition economies and ranked near the bottom of Central and Eastern European economies in terms of financing pri- vate sector activities: in 2000, credit from deposit-taking banks to the private sector was equivalent to about 9 percent of GDP. Furthermore, this figure is an overestimate of the true provision of credit because some of it reflects loans from the state banks to bankrupt former state- owned enterprises (SOEs). In 1999, Romanian banks' nonperforming loans amounted to over one-third of total loans. Average real lending rates were high, but effectively there was very little arm's length lend- ing. Financial markets were distorted by subsidized credit that pro- vided opportunities for arbitrage across different loan instruments. Following the introduction of the secured transactions reform in 2001, the impact on lending in Romania has been dramatic. In the first 18 months after the reform was implemented, there were over 324 Part II: Detailed Analysis 400,000 loans against which security interests were registered. Over 100 banks registered security interests in the filing archive. Since there are 38 licensed banks in Romania, the implication is that scores of non-Romanian banks were lending in the country against collateral and registering their security interests. Of the security interests that were registered and current in Septem- ber 2003, nearly 20 percent represented nonbank secured loans. This is especially beneficial in rural areas that do not have bank offices. In addition, the geographic coverage has been extensive. As of 2003, there were filings of security interests in 42 of Romania's 43 counties. The ratio of domestic currency lending to the private sector to GDP rose by 86 percent in the period from 2000 to September 2003, and long-term lending denominated in lei rose sharply from less than 20 percent of total credit to nearly 40 percent. The business community supports the reform enthusiastically. Many of those interviewed spoke positively of the new framework. Bankers and members of the busi- ness community were equally enthusiastic. Credit is granted not only to companies in urban areas but also in rural regions where such diverse assets as cows and tractors are taken as collateral. Given the uniqueness of the secured transactions reform, this report recommends that there be an in-depth analysis of its impact. It is important to note, however, that that the secured transactions reforms are difficult and require a great deal of attention to detail. Many attempts at secured transactions reforms in developing countries have not been successful and some are already being redone: the secured transactions reform in Ukraine initiated by the World Bank is an effort to redo the unsuc- cessful USAID reform of a few years ago. NOTES 1. This chapter was prepared by Paul Holden, Enterprise Research Insti- tute, Washington, DC. 2. Moreover, the latter two organizations do not provide a competitiveness ranking for Armenia. This is due to the lack of relevant data and survey stud- ies. A recent initiative will provide the WEF with the necessary information to include Armenia in their "Global Competitiveness Indicators" report. 3. For definitions please refer to the "Business Competitiveness Indica- tors" published semiannually by the Department of Trade and Industry in the United Kingdom (http://217.154.27.195/sd/bci/index.htm). We have not constructed business survival rates, claimant counts, or income deprivation, owing to the lack of data. 4. Business formation, employment, and average earnings figures are for 2001­04; gross value added figures are for 2000 and 2001. Strengthening Competition 325 5. The conclusions in the bulleted points that follow are a preliminary assessment; a deeper analysis based on the kind of data not yet available is required to draw firm conclusions. 6. These findings are based on evidence provided by the sample survey conducted by the NSS in 2002. 7. The dispute arose from existing inefficiencies. The mobile phone cards issued by ArmenTel were sold at a high premium on a secondary market. ArmenTel issued insufficient quantities of calling cards to meet the demand and the price was bid up to 3 or 4 times the face value of the cards. In addi- tion, ISPs (Internet service providers) were required to go through ArmenTel to connect to the Internet, which resulted in unreliable connections and restricted their ability to compete on price. Service was also limited to modem connec- tions only. ArmenTel maintained a monopoly on high-speed connections and charged prices that are 20 to 30 times more than countries with competition in the telecommunications sector. 8. ArmenTel was granted exclusive rights to a wide range of services under the 2004 license. However, Article 4.D states that under some condi- tions (which collectively may be called cases of inactiveness of the licensee in providing services), the regulator may allow another person to provide these services. It is difficult to imagine a situation where ArmenTel does not pro- vide, among other things, fixed-line, international call, and mobile services. 9. Except for small producers, which can sell at a predetermined price until 2007. 10. Another possible form of restructuring is complete vertical separation. Under this arrangement the management (and, possibly, the ownership) of facilities is fully separated from other rail functions. The main advantage of this vertical unbundling is that rail transport is placed in a similar situation as road transport, especially regarding infrastructure planning. However, this form of restructuring is currently unacceptable to the Armenian railway sector given the reasons outlined in the chapter. 11. Integrated means that the operator controls not only the infrastructure but also other railroad services. 12. The theory of contestable markets (Baumol, Panzar, and Willig 1982) helped clarify the proper definition of the natural monopoly concept. It is true that duplicating the network infrastructure in the railway industry is generally inefficient due to fixed costs that, in addition, are largely sunk. Hence, the physical network is characterized by natural monopoly and there are significant entry barriers. However, the approach may be modulated if we consider a situation in which the main activities are subject to effective competitive pressure. 13. As an example, consider the case of the Gold Mining Company. With the competitive access option implemented, if ceiling prices were set using stand-alone cost practices, the integrated carrier would have ordinary business incentives to find and cooperate with efficient participants in its businesses 326 Part II: Detailed Analysis (say, finding a new company that is cost-efficient in loading and unloading freight), as the result of this cooperation would reduce the overall cost of serving a particular shipper and leave them with extra profit to share. 14. A comprehensive analysis is contained in Chapter 13 of this book. The study includes detailed recommendations and suggestions for the liberaliza- tion of the aviation sector and its effective regulation. 15. Customs brokers were legalized only three years ago. 16. Licensing fees are US$500 per year, very high in a country where civil servants' salaries rarely exceed US$100 per month. 17. Importers are required to present original documents, including the invoice, the contract covering the import, the certificate of origin, and an export declaration from the country of origin. 18. The CPEC has investigated a number of recent cases, among which was the successful elimination of a monopoly granted by the airport operator to a taxicab company. There have also been several hearings concerning the abuse of monopoly power by ArmenTel. The company had cut off the telephone lines of some Internet service providers (ISPs) without prior notification, sus- pecting them of providing VoIP service in Armenia. In almost all cases the CPEC made decisions in favor of the ISPs and fined ArmenTel to cover the losses incurred by ISPs. 19. A four-page analysis concluded that there is competition in the distribu- tion of petroleum because of the fact that there are several wholesalers/importers within Armenia. The paper does point out that there is only one supplier/seller to the wholesalers but appeared not to view this as a problem. 20. Which attitude is the most relevant in practice is largely an empirical question. Overall insights demonstrate that difficulty in obtaining funding, rather than fear of aggressive behavior by an incumbent, is the factor which prevents firms in Armenia from entering the market. 21. The data for Armenia appear as percentages superimposed on the fig- ures. For the countries in the sample, the figure is based on the ratio of credit to the private sector from deposit money banks and other financial institu- tions. For Armenia it is the ratio of credit to the nongovernment sector (sam- ple of 47 countries). High-income countries are defined as countries where GDP per capita in 1999 was higher than US$10,000. 22. For a more comprehensive review of collateral issues, see Holden (1997). Upper-middle-income countries are defined as countries where GDP per capita in 1999 was between US$3,000 and US$10,000. Lower-middle-income countries are defined as countries where GDP per capita in 1999 was between US$1,000 and US$3,000. Low-income countries are defined as countries where GDP per capita in 1999 was lower than US$1,000 (Source: International Financial Sta- tistics Database [http://ifs.apdi.net/imf/] and the World Bank's WDI Data- base [http://web.worldbank.org/WBSITE/EXTERNAL/DATASTATISTICS/ 0,,contentMDK:20899413~menuPK:232599~pagePK:64133150~piPK: 64133175~theSitePK:239419,00.html]. Strengthening Competition 327 23. Some advocate "relationship lending" in terms of which banks make loans on the basis of the analysis of business plans and the borrower's his- tory as an entrepreneur. This is unlikely to happen. First, most lending to businesses in the United States is secured by collateral. Second, the skills available for drawing up business plans are scarce. Third, those who have run businesses successfully are often the entrenched interests in Armenia, so that loans go to those who have tight control of the business sector. Chapter 11 Real and Formal Ease of Doing Business in Armenia in Comparative Perspective: Implications for Regulatory Reforms "Those who wish to offer their countries the chance to become a mod- ern society must direct their efforts not only toward eliminating cer- tain conditions--the barriers of access to formal activity, the costs of remaining legal, the discrimination in capital markets, the state corpora- tions--but also toward creating the basic institutions on which efficiency, social peace, and spontaneous cooperation are based in more advanced nations." (De Soto 1999, p. 259) INTRODUCTION Well-developed, market-supporting institutions, the contestability of domestic markets, the capacity of a government to enforce the rule of law, and macro-stability are all crucial to establishing an environment friendly to conducting business operations.1 But so are regulations and other conditions influencing the transaction costs of business opera- tions. Regulations, while important, do not have to be unnecessarily burdensome or costly, thereby deterring the entry of new businesses or undermining the competitiveness of existing ones. The high "hassle" cost of conducting business provides a strong disincentive to establishing new firms and encourages existing firms to slip into the informal economy. In a similar way, burdensome customs procedures, the low efficiency of ports, high transportation costs due to regulatory arrangements, and high costs of other services raise transaction costs and undermine the competitiveness of domestic producers in international markets. To successfully take advantage of growth and employment opportunities in domestic dimensions, the hassle cost of conducting business has to be reduced in both domestic 329 330 Part II: Detailed Analysis and foreign dimensions. Better facilitation measures for foreign trade will be of little relevance if the domestic environment discourages the entry of new businesses and encourages the exit of firms either per- manently or to the informal economy. Similarly, the limited capacity of firms to recover payments due from other firms discourages increased ties and limits the intensity of business activity, with negative conse- quences for domestic production and exports. This analysis focuses on the domestic dimension of conducting business using the results of the World Bank's Doing Business sur- vey (http://www.doingbusiness.org). In contrast to various other assessments of the quality of governance,2 the Doing Business sur- vey captures the tangible legal components of the investment climate together with their cost-burden and time-burden imposed on busi- nesses across 145 countries (Djankov et al. 2000). The World Bank's Cost of Doing Business comparative survey (World Bank 2005) pro- vides a useful diagnostic tool for conducting an initial assessment of the quality of the business environment in Armenia. It provides data on the tangible components shaping the cost of conducting busi- ness in Armenia and allows casting them in a comparative perspec- tive. The survey covers the following seven areas: starting business, labor market flexibility, registering property, contract enforcement, bankruptcy, protection of investors, and getting credit. Each area has several different indicators ranging from 1 (protection of investors) to 5 (getting credit and labor market flexibility). These are not firms but local professionals dealing with issues related to each of these areas that have been surveyed in respective countries in both large cities and rural areas. The Cost of Doing Business's advantage over surveying firms is obvious. Professionals deal with respective issues on a day-to-day basis, whereas firms confront the issues only to the extent that the issues directly affect them. The former have an overall view, whereas the latter do not. The procedure used to assess the ease of doing business in Armenia consists of two stages. First, we identify Armenia's relative position among transition economies in terms of an overall, formal ease of doing business. We identify best-practice indicators for each "doing business" area--not worldwide but among transition countries. Indi- cators for countries are expressed in terms of percent of the best prac- tice set at 100. These are demanding benchmarks, as two transition countries--Lithuania and the Slovak Republic--are among the top 20 economies in the world on ease of doing business. We average the benchmarks for each of the seven areas covered in the Doing Business survey to obtain an overall assessment of the ease of doing business for each country. Second, in order to capture the full extent to which the business environment is friendly or unfriendly, we bring to the fore the size Real and Formal Ease of Doing Business in Armenia 331 of the informal economy. The rationale is straightforward. The use of Cost of Doing Business indicators allows assessing the extent to which a country deviates from the best practice on each indicator within each area in terms of business laws and regulations. It does not address the quality of the state administration. Neither does it assess hassles associated with tax policies, tax and customs administration, frequency of inspections, cost of licenses, and so forth. While there are many explanations why businesses decide not to go formal, three reasons stand out: low chances of getting credit, fear of predatory administration and discreetly applied taxes, and lack of trust in the state's capacity to enforce contracts. While the survey captures both getting credit and contract enforcement, it does not capture informal administrative interventions. Moreover, the overall ranking procedure assigns the same weight to all areas of doing business, and within each area the procedure assumes that each indicator implies similar levels of ease or hassle. Regulations may be very liberal on all indica- tors within an area except one: as a result, a country may have a very high overall score in an area, while in reality one barrier makes other friendly regulations completely irrelevant. The size of the informal sector provides indirect information on the weight of formal barriers as well as on variables not covered by the Doing Business survey. We assume that the sector provides infor- mation equal in terms of weight to the overall formal ease of doing business. Therefore, we take the average of performance in terms of the informal economy and ease of doing business as an indicator of the quality of the business environment. This index is referred to as the "revealed" or real ease of doing business. Detailed analyses of Armenia's performance on each indicator in each area follow the logic presented above, but with a twist. First, strengths and weaknesses for each area are identified and assessed in the context of other transition countries. Second, since there appears to be a gap between Armenia's relatively good performance in the formal ease of doing business and the large size of the informal sector, the analysis proceeds to identify factors that discourage firms to go formal. OVERALL EASE OF DOING BUSINESS IN ARMENIA AND OTHER TRANSITION COUNTRIES Table 11.1 presents the ranking of CIS and CEEC-10 economies in terms of the ease of doing business.3 For each area, the values of indi- cators were normalized in terms of best practice in CIS and CEEC-10 economies. For instance, two areas of flexibility of employment are taken into account: (i) the Rigidity of Employment Index, which is an average of three indices (hiring, hours, and firing); and (ii) firing 332 Part II: Detailed Analysis costs, expressed in the number of weeks that an employer has to pay wages. Every indicator for each country is expressed in values ranging between 100 (best practice among CIS and CEEC-10 economies) and a lower value that is larger than zero. However, the value of the indi- cator is set at zero--representing the worst possible practice--when what might be said to be the law of jungle prevails, for example, in the absence of any laws or regulations directed at protecting the interests of domestic or foreign investors. The aggregate index of the ease of doing business in the post-communist world is the average of average values for each area. In terms of the overall formal ease of doing business, Armenia ranks sixth among 20 transition economies. Doing business across all areas is on average easier than in three new EU member states (Hun- gary, Poland, and Slovenia) and easier than in all other CIS countries as well as in Bulgaria and Romania, the newest members of the EU. Only three Baltic states (Estonia, Latvia, and Lithuania), the Czech Republic, and the Slovak Republic--all of them with strong reform credentials--score higher than Armenia. Within areas, Romania ranks highest in starting a business, the Slovak Republic in flexibility of the labor market, Lithuania in registering property and enforcing con- tracts, Latvia in closing a business, the Slovak Republic together with the Czech Republic and Lithuania in protecting investors, and the Czech Republic in access to credit (the high scores are marked in italics in Table 11.1). Armenia owes its elevated ranking to the greater flexibility of its labor markets as compared to those in other transition economies; but broader comparisons suggest that labor markets are fairly rigid, par- ticularly taking account of the size of compulsory severance payments. It owes its ranking also to simplicity and effectiveness in enforcing contracts and effective procedures in closing a business. Armenia's weakest areas are registering property and access to credit. Overall, however, even in the most streamlined and reformed areas, the level is well below the best practices in transition economies. Armenia's scores do not set it apart from countries ranked higher in ease of doing business, as even countries with the best indicators for any given area have a long way to go to catch up with the best practice. As can be seen from the last line of Table 11.1, the distance is particularly large for labor market flexibility and registering prop- erty. While the comparison involves "ideal types," (that is, nonexistent economies with the best set of indicators either region-wide or world- wide), it nonetheless demonstrates that even the best practice in the region is far away from the best practice in the world. The obtained ranking in terms of overall ease of doing business is surprising in two respects. First, considering the pressure of accession to the EU and the achieved degree of harmonization with the acquis Table 11.1 Ranking of CIS and CEEC-10 Economies by Ease of Doing Business in 2005 Starting Flexibility of Registering Enforcing Closing Protecting Getting Ease of doing Country business employment property contracts business investors credit business Real Armenia 47 64 38 65 68 50 37 53 Azerbaijan 44 33 29 56 43 33 24 38 and Belarus 18 50 48 54 44 17 44 39 Bulgaria 35 46 19 48 41 33 69 42 Formal Czech Republic 33 56 28 72 18 100 89 57 Estonia 40 37 40 83 45 67 63 53 Georgia 39 51 18 54 53 83 56 50 Ease Hungary 34 38 27 76 36 83 40 48 Kazakhstan 41 69 18 60 24 83 23 45 of Kyrgyz Republic 70 54 22 28 53 50 31 44 Doing Latvia 48 30 15 77 100 83 70 61 Lithuania 58 37 74 86 68 100 40 66 Business Moldova 33 50 26 51 41 50 36 41 Poland 32 49 19 51 60 67 57 48 Romania 79 17 17 51 27 33 53 39 Russian Federation 43 69 21 62 77 50 13 48 in Slovak Republic 39 100 27 49 31 100 54 57 Armenia Slovenia 28 28 20 45 27 67 48 38 Ukraine 27 17 14 65 32 50 28 33 Uzbekistan 33 39 10 45 47 67 39 40 Source: Own calculations based on data in WB 2005. 333 334 Part II: Detailed Analysis communautaire, that is, the body of laws governing relations within the European Union communities, one would expect the CEEC-10, in particular the 8 new EU member states, to dominate the top 10 countries in terms of the formal ease of doing business. To a large extent this is the case, as 5 new members top the list. But two CIS members--Armenia and Georgia--outperformed Poland. Further- more, another CIS country, Russia, also appears to have a friendlier regulatory framework for conducting business than two other new EU members, Hungary and Slovenia. The latter appears to have the most burdensome regulatory regime next to Ukraine, which is at the bottom of the list (Table 11.1). Bulgaria and Romania are relatively low in ranking in the formal ease of doing business, although they have already signed a protocol accession to the EU indicating the European Commission's seal of approval of their institutional and policy readi- ness to meet the requirements of membership scheduled for 2007. In addition to the countries mentioned above, Bulgaria ranks thirteenth, below Kazakhstan and the Kyrgyz Republic, while Romania ranks sixteenth, below Uzbekistan and Moldova. In contrast to most CIS economies, all CEEC-10 economies appear to score high on indicators describing the quality of the legal infra- structure of financial markets, in particular the bankruptcy laws and collateral as well as market information on the quality of borrowers. They compare favorably with international standards. But some of them, most notably Slovenia, clearly fail on other counts. Second, the relatively high position of Uzbekistan, a country that has made little progress in establishing market-supporting institutions, comes as another surprise. Its high position is due to the laws protect- ing investors and the ease of closing a business. These two areas alone are unlikely to encourage business start-ups and new investments in the presence of other impediments. Both surprises stem in large part from the fact that the Doing Busi- ness survey takes into account selected formal ingredients of the busi- ness climate. Furthermore, the procedure that we use to rank econo- mies in terms of the overall formal ease of doing business assigns equal weight to areas and indicators within areas. For instance, a low number of procedures would put a country higher up, which might not be offset sufficiently by a lengthy period needed to complete a procedure.4 As a result, the ranking may be higher than the actual ease of completing a required procedure. "REVEALED" EASE OF DOING BUSINESS The ease of doing business indicators for CIS/CEEC countries can- not capture each important ingredient determining the quality of the Real and Formal Ease of Doing Business in Armenia 335 business environment, simply because they are not designed to do so. Two examples illustrate what is left out.5 First, one of the "real world" tests of the quality of the regulatory regime is the size of the informal sector. Weak enforcement of property rights and difficulties in access to bank credits, combined with a capri- cious, corrupt, and predatory administration, keep businesses in the shadow economy. This implies that in countries with a higher value of the aggregate indicator for ease of doing business, the informal economy should be smaller, and vice versa. For instance, international evidence suggests that labor market rigidity is closely associated with the size of the informal economy (that is, the less flexible labor arrange- ments are the more frequent is exit to the informal sector). Yet the value of the correlation coefficient between values of the overall ease of doing business indicators and the size of the infor- mal economy in terms of gross national income (GNI) is, as expected, negative, but still low at 38 percent. Upon closer examination, this finding should not come as a surprise. Consider for instance a simple, relatively business friendly regulatory regime deficient in one dimension, namely, very difficult access to credit. This alone might discourage an entrepreneur from registering a small firm: while there is no reward in establishing credit, there is a penalty in terms of formal payments. Considerations of tax evasion and avoiding other possible nuisances from the state administration may prevail over going legal. This seems to be the case with many CIS countries, especially Russia and Armenia, both of which score high in other areas of doing business but rank relatively low in the area of getting credit. In both countries financial markets function poorly, nullifying one of the main reasons for going formal. Poor mar- ket functioning does not provide Armenian businesses with incentives to formalize, particularly since this situation also allows petty official- dom to target formal businesses with inspections that have payoffs as their aim. On the other hand, easy access to credit provides an incentive to move to or stay in the formal sector. For an entrepreneur seeking access to financing as well as to the system of formal contracting and legal protection in business transactions, credit alone may override concerns about other formal hassles associated with doing business. Access to credit probably explains the relatively low size of the infor- mal economy in Poland or Slovenia, despite significant barriers to doing business. Second, one would also expect the ease of doing business to be positively correlated with the falling incidence of corruption. But again this is not the case, at least when using the values of the Cor- ruption Perception Index (CPI), annually estimated by Transparency International surveys, as a measure of the incidence of corruption in a 336 Part II: Detailed Analysis country. The correlation with ease of doing business values is positive but low at 33 percent. Since corruption is highly correlated with the size of the informal economy, the use of the latter takes care of issues that are left out in the overall ease of doing business indicator. The ranking in terms of the values of CPI, averaged over 2000­04, does not widely diverge from that generated by the size of the informal economy. Countries perceived as having a lower incidence of corruption (that is, a higher value of CPI) tend also to have a smaller informal sector in terms of GDP. The correlation between corruption and the informal sector is 67 percent. Excluding the biggest spoiler, Belarus, the correlation goes up to 79 percent.6 Thus, a comprehensive assessment of the quality of the business climate should take into account both the formal ease of doing busi- ness and the size of the informal sector. Since the size of the informal sector is a less subjective measure than a CPI and both are strongly correlated, it is sufficient to incorporate information about the infor- mal economy into the composite index of ease of doing business. The proposed measure, which is thereafter referred to as the "revealed" overall ease of doing business,7 assigns equal weight to formal ease and the size of the of informal economy. The size is expressed in terms of GNI and is normalized along similar lines as indicators of ease of doing business (Table 11.2). The values of the informal sector size are expressed on a scale from 1­100, with 100 assigned to the lowest value of the informal sector in CIS/CEEC coun- tries. The Slovak Republic has the lowest size and therefore the value taken into account in calculating the "revealed ease" is 0.5 of 100. The lowest normalized level of this indicator is 28 for Georgia, with the highest informal economy of 67.3 percent in 2003. Except for Uzbekistan's high position, where the relatively low size of the informal economy may be due to strong state controls, the rela- tive rankings reflect countries' progress in building an institutional environment supporting private business activity. New EU members occupy the top 8 spots, with Slovenia moving up 11 spots compared to the overall ease of doing business indicator. On the other hand, several countries have dropped in overall ranking, with Georgia going down 10 spots, followed by Russia and Armenia. Thus, incorporating the size of the informal sector would appear to provide a better indication of the business-friendly conditions of the existing regulatory arrangements than an indicator of an overall ease of doing business. As demonstrated above, the informal sector and unfriendly business regulation usually go together, with some excep- tions. For reasons discussed earlier, the regulatory regime in Slovenia, despite the apparently heavy burden it imposes on businesses, does not encourage entrepreneurs to go underground. However, appar- Table 11.2 Size of the Informal Economy and "Revealed" Ease of Doing Business "Revealed" ease Overall ease Difference between Informal economy of doing of doing revealed and Memorandum: Country percent of 2003 GNI business indicator business indicator overall indicator CPI, 2000­04 Real Armenia 46.3 46.8 6 5 3.0 Azerbaijan 60.1 34.5 18 2 1.9 and Belarus 48.1 39.2 17 1 4.1 Bulgaria 36.9 46.4 13 ­1 4.0 Formal Czech Rep. 19.1 77.8 4 ­2 3.9 Estonia n/a 74.0 5 ­2 5.7 Georgia 67.3 39.3 7 10 2.1 Ease Hungary 25.1 61.5 10 ­5 4.9 Kazakhstan 43.2 44.6 11 3 2.3 of Kyrgyz Rep. 39.8 45.8 12 1 2.2 Doing Latvia 39.9 53.9 2 5 3.8 Lithuania 30.3 64.2 1 3 4.7 Business Moldova 45.1 41.4 14 2 2.3 Poland 27.6 58.2 8 ­2 3.8 Romania 34.4 47.2 16 ­6 2.7 Russian Federation 46.1 44.4 9 6 2.7 in Slovak Rep. 18.9 78.5 3 ­2 3.8 Armenia Slovenia 27.1 53.7 19 ­11 6.0 Ukraine 52.2 34.6 20 ­1 2.3 Uzbekistan 34.1 47.7 15 ­6 2.5 Source: Authors' calculations based on data in WB 2005 and Web site of Transparency International (http://www.transparency.org/). 337 Note: CPI is the average for 2000­04. 1 = maximum incidence of corruption; 100 = minimum incidence of corruption. 338 Part II: Detailed Analysis ently friendlier regimes in Armenia, Georgia, and Russia are not suf- ficient to retain businesses in the formal sector, to the detriment of economic growth and productivity. International experience suggests that productivity levels are two to four times lower in the informal sector than in the formal economy. When the size of the informal economy does not move in tandem with the difficulty of doing business, it suggests that other ingredients of the business climate that are not captured by the Doing Business survey effectively discourage potential entrepreneurs from enter- ing the formal economy. These other "negative" factors are strongly present in Armenia, Georgia, and Russia. Negative factors usually stem from capricious and predatory administrations whose interventions effectively weaken the poten- tially positive impact of regulatory reforms. They include excessive costs of transparency, burdensome tax regulations, and high, unstable tax rates. These factors may raise the hassle cost of doing business to levels not tolerated by most businesses. However, these explanations, together with the incidence of corruption, are generic answers, and more detailed analysis is needed in the particular country context. WEAKNESSES AND STRENGTHS OF THE REGULATORY REGIME How does Armenia score in various areas of doing business? We bench- mark the values of the respective indicators for Armenia against those for CEEC-8 countries,8 CIS economies, Russia, New Zealand, and best practice across the globe. The rationale behind the choice of CEEC-8, CIS economies, and Russia as comparators is straightforward. Except for Slovenia, they all share the legacy of central planning or, more precisely, of the overregulation and micromanagement of the economy by the administration. Therefore they provide a meaningful frame of reference for assessing the progress in establishing a business-friendly environment. Furthermore, Armenia potentially competes with them for foreign investments while it does not benefit from the advantage of its geographical location close to the core of the global economy. Russia has the least onerous regulatory environment among CIS econ- omies across all dimensions of doing business, whereas New Zealand scores highest in the world in terms of ease of doing business. Table 11.3 also contains information about the best values worldwide for each indicator. Reviewing the data in Table 11.3 leads to the following general observation: the regulatory burden of conducting business in Armenia is relatively low, especially when benchmarked against CIS and CEEC-8 economies. Real and Formal Ease of Doing Business in Armenia 339 On a positive note, Armenia fares quite well in several areas of doing business. Entry is easy. So is exit as well as transferring property rights. It is easy to start a business in Armenia. It costs only US$64 in administrative fees, requires a minimum capital of US$44 that the entrepreneur has to deposit before registration kicks in, and takes 25 days. Contrast these entry conditions with US$116, US$819, and 32 days in a median CIS economy or US$142, US$3,199, and 46 days in a median CEEC-8 economy. Although the difference in favor of Armenia is smaller as far as costs are concerned, when expressed in relation to GDP per capita it remains considerable. Armenia does not make it difficult--relative to CIS and CEEC-8 comparators--for investors to register property (that is, to transfer the property title from a seller to a buyer). It is a relatively simple operation involving four procedures and costing less than 1 percent of the value of the property. It takes less time (18 days) than in compara- tor countries, excluding the world's least repressive countries toward businesses. Neither do Armenian regulations create barriers for businesses that go bankrupt and are forced to exit. Claimants recover 39.6 percent of the amount they are owed. This is less than in Russia but almost the same as in a median CEEC-8 country and considerably more than in median CIS. It also takes less time to resolve foreclosure or bank- ruptcy and the cost of closing a business is a smaller percentage of the value of a troubled business. The cost (4 percent) is the same as in Russia but the time (1.9 years) is longer. Since the worst indicator does the most to discourage business activ- ity under most circumstances, the following question is pertinent to an overall assessment of the regulatory and policy hassle felt by busi- nesses: which aspects of Armenia's regulatory environment are the most burdensome for private firms? While overall in each area Arme- nia scores high compared to its peers from CIS and CEEC-8 countries, within some areas the indicators point to weaknesses. These areas are labor markets (firing and hiring), access to credit, and contract enforcement. As for contract enforcement, it takes more money (18 percent of the debt as compared with 14 percent), but other indicators favor Armenia. In a very important dimension of labor markets, Armenia's busi- ness environment has significant weaknesses even when benchmarked against only other CIS countries. The high value of the "difficulty to fire index," combined with an obligation to pay the wages of a fired worker over a period of 30 weeks, significantly reduces the flexibility of Armenia's labor market. More favorable values for the remaining two indicators do little to lessen constraints on hiring and firing in response to business conditions. 340 Table 11.3 Doing Business in Armenia and Selected Comparators, 2004 Part Best practice CIS-10 CEEC-8 in CEEC New Best II: Area Indicator Armenia median Russia median and CIS Zealand practice Detailed Starting a Number of procedures 10 10 9 9 5 2 2 business Time (days) 25 32 36 46 18 12 2 Analysis Cost (% of income per capita) 7.0 14.2 6.7 12.0 3.7 0.2 0.0 Minimum capital (% of income per capita) 4.5 23.3 5.6 47.9 0.0 0.0 0.0 Hiring and Difficulty of hiring index firing workers (0­100) 33 33 0 33 0 11 0 Rigidity of hours index (0­100) 40 60 60 60 20 0 0 Difficulty of firing index (0­100) 10 60 20 60 10 10 0 Rigidity of employment index (0­100) 28 44 27 44 10 7 0 Firing costs (severance payments in weeks of wages) 30 21 17 21 17 0 0 Registering Number of procedures 4 7 5 5 3 2 1 property Time (days) 18 71 354 72 3 2 1 Cost (% of property value) 0.9 2.2 9.5 2.1 0.5 0.2 0.0 Table 11.3 (continued) Best practice CIS-10 CEEC-8 in CEEC New Best Area Indicator Armenia median Russia median and CIS Zealand practice Real Enforcing Number of procedures 24 29 29 24 17 19 11 contracts Time (days) 195 324 395 333 150 50 27 and Cost (% of debt) 17.8 18.0 9.5 10.8 8.1 4.8 4.2 Formal Closing a Time of insolvency (years) 1.9 3.8 1.5 2.5 1.1 2.0 0.41 business Cost (% of estate) 4.0 7.4 4.0 18 4.0 4.0 1.0 Recovery rate (cents on Ease the US$) 39.6 5.3 48.4 39.8 85.0 71.4 92.0 of Protecting Investors disclosure Doing investors index (0­7) 3 3 3 5 6 5 7 Getting credit Cost to create collateral Business (% income per capita) 0.9 3.6 11.6 3.7 0.6 0.0 0.0 Legal rights index (0­10) 4 5.0 3 6 9 9 10 Credit information index (0­6) .. 0.0 0 4 5 5 6 in Public registry coverage Armenia (per 1,000 adults) .. 0.0 0 6 44 0 637 Private bureau coverage (per 1,000 adults) 0 0.0 0 17 380 978 1,000 Source: Data from World Bank (2005). 341 .. Negligible. 342 Part II: Detailed Analysis Research shows the crucial importance of labor market flexibility for both reducing unemployment and boosting investment. While the relationship between low unemployment rates and high flexibility is empirically firmly established, the impact of labor market flexibility on FDI appears to be significant as well. A recent study (Javorcik and Spatareanu 2004) using the labor market indicators from the Doing Business database for 2002 shows that, all else being equal, flexibility in labor markets has a significant impact on FDI flows. For instance, if the flexibility of the host country labor market increases from the level of the Slovak Republic (inflexible prior to reforms in 2003) to the level of Hungary (flexible), the volume of FDI goes up by between 14 and 18 percent. Moreover, in line with intuition, FDI in service sectors appears to be more sensitive to labor regulations than investment into manufacturing. Armenia also has a relatively low score in terms of investors' pro- tection, with the disclosure index of 3 putting it on a par with Russia. The absence of regulations compelling disclosure has several negative consequences: investment is lower as potential investors fear expro- priation, the stock market is undercapitalized, economic growth is lower than under full disclosure, and returns from investment are lower (World Bank 2005, pp. 56­57). But more powerful financial constraints to doing business relate to other legal underpinnings of financial markets in Armenia. Although the cost of creating collateral is very low by both regional and interna- tional standards and the legal rights index is within a median range, this should not necessarily imply easy conditions for getting credit. The contrary is the case, as the difficulties faced by firms in obtain- ing credit in Armenia clearly demonstrate. The absence of the insti- tutionalized information on borrowers that should be available to the banking sector curtails lending and also offsets the potentially positive impact of the relatively low costs of creating collateral and the decent standards of protection of creditor rights in bankruptcy. The lack of market information on the quality of borrowers, com- bined with the absence of a secured transactions framework, is a bind- ing constraint on lending. Furthermore, movable property cannot be used effectively as collateral to secure loans in Armenia, as repos- session is time consuming and costly. Neither can collateral-based lending function effectively, as Armenia does not have an effective framework for creating and enforcing claims--that is, a framework for secured transactions is not yet in place (Holden and Sahakyan 2004). The effects of the inadequacies of the collateral framework extend throughout the economy, with the result that banks correctly perceive lending as extremely risky. Real and Formal Ease of Doing Business in Armenia 343 The gap between Armenia and the median of comparators is lower in registering property: the number of procedures is only slightly higher (4 versus 3) but it takes significantly more time to complete the process (18 days versus 3 days). This is also the case in closing a business. The time of insolvency is longer (1.9 years versus 1.2 years), although it is lower than in New Zealand (2 years), with the best international practice at 0.41 years. The recovery rate of 39.6 cents on the U.S. dollar (that is, the percent of what creditors collect on their debt) is below Lithuania's level, but similar to the CEEC-8 median of 39.8. An indicator on starting a business is lagging behind comparator values; the indicator is higher than in Lithuania but lower than the median value for the CEEC-8. Although Armenia appears to be a top performer among CIS coun- tries, this is not sufficient. An important consideration is that Armenia has to neutralize the negative effects of its remoteness to the centers of gravity of global markets and of its being landlocked. The frame of reference for Armenia's regulatory reforms should be at least the best practice in the CIS/CEEC-10 region, if not the best international practice. Here, on both counts, the differences are huge, even though Armenia fares quite well overall in relation to other former centrally planned economies. THE COEXISTENCE OF FORMAL EASE OF DOING BUSINESS AND THE INFORMAL ECONOMY Why does such a large informal economy coexist with relatively high formal ease of doing business? The answer may be found in identify- ing the specific factors that raise the cost of doing business above the "formal" levels, namely those factors identified in the Doing Business survey. In other words, what are the main ingredients contributing to the fact that the actual costs exceed the formal costs of compliance? Interviews and surveys covering various areas of the business envi- ronment paint a picture of general dissatisfaction on the part of busi- nesses with the quality of the regulations, administrative requirements, and bureaucratic behavior of state agencies, despite some improve- ments over the last three years. The reasons for this assessment are manifold. First, difficulties encountered in interaction with tax adminis- tration, compounded by unstable and frequently changed rules and tax rates, are cited as a major barrier to conducting business in Armenia. This is perceived as a bigger problem than corruption by both local- and foreign-owned firms, albeit for local firms corruption continues to be regarded as a significant barrier.9 Administrative Regulatory Cost 344 Part II: Detailed Analysis Surveys (RCSs), which have been conducted in Armenia on an annual basis since 2000, persistently identify tax administration as the greatest obstacle to conducting business. For instance, in both 2003 and 2004 more than 80 percent of respondents identified tax administration as the largest obstacle. The view appears to be shared by foreign-owned firms to the tune of 90 percent of all respondents (FIAS 2003). Further- more, foreign-owned businesses complained that the tax burden in 2003 considerably increased compared to 1999. Uncertainty associated with frequent changes in policy was flagged by around three-fourths of RCS respondents. The second barrier relates to the costs that businesses have to incur in order to meet the requirements of administrative regulations. According to the RCS, the costs remain huge and there has been no major improvement over the past several years. Most costs are not related to formal and informal payments but to the imposed burden of time and resources that firms must allocate to assure regulatory compliance. The third barrier is the "irregular" payments extracted by customs and the time consuming procedures. Customs clearance procedures are applied with equal zeal not only to imports but also to exports. The time needed to complete customs clearance as well as the amount of effort undertaken by a firm to deal with customs are shocking, especially for exports. Furthermore, traders are charged by customs for the storage of the held goods. Such conditions add to the overall high cost of conducting foreign trade transactions in Armenia and corroborate businesses' complaints about nontransparent procedures and nontariff trade barriers. In consequence, it appears that the most obvious benefits of the computerization of customs services (that is, reducing the release time of consignments and slashing the documentation or bureaucratic bur- den put on a trader or customs broker) have so far failed to material- ize. Neither traders nor customs brokers have access to the system. The old practices of bureaucratic delays fueling corruption combined with lack of capacity in customs administration continue. Finally, although Armenia is no exception to the generally poor record of most transition countries in establishing the rule of law and strengthening the courts, the situation has improved on several impor- tant counts.10 Armenia has made large strides in judicial reforms and is "the only country where significantly more respondents viewed the courts as fairer in 2002 than in 1999" (Anderson, Bernstein, and Gray 2005). Considering, however, that in 2000 only 2 percent of firms sur- veyed viewed judges as honest (CES 2000), the improvement should be measured from the low starting point. Fairness, however, does not exhaust all of the dimensions related to efficiently functioning courts. Real and Formal Ease of Doing Business in Armenia 345 Other important dimensions from the point of view of conduct- ing business include the perceived honesty or dishonesty of judges, the capacity to implement enacted commercial laws, and the ability to adjudicate disputes in an efficient and timely manner. It appears that firms regard Armenia's courts as relatively honest. In response to a question about the frequency of unofficial payments when deal- ing with the courts, firms surveyed under Business Environment and Enterprise Performance Surveys (BEEPS) identified the courts as not extracting bribes on a significant scale. The percent of firms nega- tively assessing courts on this dimension was among the lowest in the region and on a par with Slovenia, Estonia, and Lithuania. Moreover, Armenia, together with Lithuania, recorded the largest improvement among 26 transition economies between 1999 and 2002. Although most respondents perceive the courts in transition countries as nei- ther honest nor fair, the percent of those who do not share this view is comparatively high at 30 percent, putting Armenia in the same group as Croatia and Latvia. Like most other transition economies, Armenia has experienced the growing gap between the enacted legislation that has laid the ground for a market economy and the capacity of courts to implement new commercial laws. This gap between the extent of commercial legis- lation existing in 1999 (the so-called "legal extensiveness") and the degree to which it was being implemented at that time (that is, the "legal effectiveness") was particularly high for Armenia. It was well above the level in other transition countries (EBRD 1999). Although this "implementation gap" persisted in 2002, it fell drastically, reach- ing the level registered in the Czech Republic. A very important dimension from the perspective of firms is the ability of courts to adjudicate disputes in an efficient and timely man- ner, as this reduces uncertainty in actual or potential business deals. The BEEPS surveys asked respondents about the fairness, honesty, affordability, ability to enforce decisions, and speed of the courts. Armenian courts had improved dramatically in terms of speed. In fact, Armenia, together with Hungary, Macedonia, Latvia, Lithuania, and the Czech Republic, showed the largest improvement in the per- ceived speediness of courts between 1999 and 2002. While in 1999 only 10 percent of firms viewed courts as quick, in Armenia this propor- tion rose to 25 percent of respondents, comparable to Hungary. This should rise further once the case management and court administra- tion systems currently being developed are put in place to help fill the statistical and knowledge gap that has been reducing the effectiveness of the courts in Armenia. According to firms participating in the BEEPS, Armenian courts are not only quick but also affordable by CEEC/CIS standards. 346 Part II: Detailed Analysis To sum up, the reasons for the large size of the informal sector in Armenia stem mostly from the high regulatory compliance costs that are due to official and unofficial payments, the instabilities in the state policies toward the business sector, and predatory tax and customs administrations, but not from weaknesses in the judicial system. While efforts designed to reform the judiciary should continue, improving tax and customs administration together with injecting stability into business regulations and policies stand out as major "other" ingredi- ents that account for the gap between the formal and the revealed real ease of doing business in Armenia. CONCLUSIONS AND POLICY RECOMMENDATIONS What does this imply for Armenian reforms? First of all, these "other ingredients" that allow the state administration (for example, tax, cus- toms, and sanitary administrations as well as building inspections) to harass businesses need to be dealt with. Beyond the reform of tax and customs administration and the continuing work on simplifying the state­business interface, other immediate measures might be imple- mented. These would include fixing tax rates for a specified period of time in a manner assuring the credibility of the government's com- mitment; limiting the number of inspections, audits, and fact-finding visits allowed over a specified period of time;11 and reviewing and simplifying customs clearance procedures. Although "other ingredients" should be at the top of the reform agenda, this should not imply a lack of urgency in bringing the dis- cussed areas of doing business in line with best practice. Armenia has to neutralize the negative effects of its remoteness to the centers of gravity of global markets and of being landlocked. While the authorities should aspire to move toward the best prac- tice in each area and in each indicator, some weaknesses could be overcome by instituting "quick fixes." Such fixes assume that political opposition to the loss of rents usually associated with opaque regula- tions could be overcome. The fixes might include the following meas- ures in the areas listed below: Getting credit. Establish market information on the quality of borrowers and undertake other measures that facilitate access to financing. This would involve, in particular, establishing a framework for secured transactions to ensure conditions for col- lateral-based borrowing.12 Reforming the labor market. Review and change labor regula- tions with the goal of removing those that do not allow for flex- ible and open (in terms of tasks performed) term contracts; those Real and Formal Ease of Doing Business in Armenia 347 that limit the number of work hours or the number of shifts; and those that compel an entrepreneur to pay a severance package currently comprising the equivalent of 30 weeks' wages. In par- ticular, the last should be drastically reduced, if not zeroed, as is the case in New Zealand. Starting a business. Streamline the number of procedures to two, remove the capital requirement, and lower fees and other admin- istrative charges. Although these significant capital requirements are often justified as protection against damage from failing busi- nesses, countries with the friendliest conditions for business startups impose very small costs, if any. The reasoning behind simplicity and low administrative costs is simple: to encourage new business activity and ensure that it takes place within the realms of the formal economy. Registering property. Reduce the time needed to register prop- erty. Ease of registering property is a necessary condition, albeit not sufficient, to bring assets into the formal sector, which, in turn, allows using the assets to obtain financing. Protecting investors. Introduce laws covering the remaining areas of disclosure of financial information and ownership that will protect investors and shareholders against fraud. Armenia has three out of seven relevant regulations. However, their posi- tive impact hinges critically on improvements in assuring the legal protection of investors. Enforcing contracts. Improving contract enforcement is a long- term project involving the review of judicial procedures and the organization of courts. Given its importance, which is growing with economic development, it should be a constant fixture on the government's reform agenda. Although the hassle cost of doing business has been significantly reduced, the challenge ahead is to bring regulations in line with the best international practice. Because of its remote location from major world markets, Armenia needs to remove all administrative barriers that unnecessarily raise transaction costs including the costs of entry and exit as well as the costs of conducting business operations. NOTES 1. This chapter was prepared by Bartek Kaminski, University of Maryland, College Park, and the World Bank. 2. Transparency International CPI scores, EBRD scores of the progress in transition, and the World Bank governance indicators capture either the per- ception of investors (transparency) of the extent of graft in the country or the 348 Part II: Detailed Analysis relative performance of a country in terms of implemented reforms (EBRD Transition Reports) or an overall assessment of a country's capacity to govern (Kaufmann, Kraay and Mastruzzi 2003). 3. The CEEC-10 includes eight new EU members (the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, the Slovak Republic, and Slov- enia) and two prospective 2007 members, Bulgaria and Romania. The CIS includes Armenia, Azerbaijan, Belarus, Georgia, Kazakhstan, the Kyrgyz Republic, Moldova, the Russian Federation, Ukraine, and Uzbekistan. The Doing Business survey does not cover two remaining CIS members, Tajikistan and Turkmenistan. 4. This should not suggest that streamlining procedures is an exercise in futility. As long as it is accompanied by fewer documents and permissions, it may ultimately help reduce time spent on procedures and will certainly curtail the potential for extracting bribes. 5. For a more detailed discussion, see Kaminski (2005). 6. Countries that "spoil" to the largest extent an otherwise almost perfect picture are Belarus, Slovenia, and the Slovak Republic. The first two have large informal economies and a low perception of corruption, whereas the Slo- vak Republic has higher corruption in relation to its small informal sector. 7. It is "revealed" because it takes into account the extent to which busi- nesses express a preference to opt out from the official economy or to stay in it. 8. The CEEC-8 countries are the Czech Republic, Estonia, Latvia, Lithuania, Hungary, Poland, Slovenia, and the Slovak Republic. 9. According to the 2004 RCS, 71 percent of respondents identified cor- ruption as a persistent problem, up from 66 percent over the previous year. Interestingly, corruption appears to be much less of a problem for foreign- owned firms than for domestic companies (FIAS 2003). 10. According to the World Bank study reviewing progress in judicial reform in transition economies, progress has been much weaker in all tran- sition countries than in other areas of institutional reform since 1990. Fur- thermore, firms' perceptions of the legal and judicial systems in transition countries are worse than comparable perceptions in most other regions of the world, according to a recent world-wide survey of business executives (Anderson, Bernstein, and Gray 2005). 11. Although tax officials can conduct only one audit per year, they are allowed to undertake as many "fact-finding" visits, not classified as an audit, as they wish. Furthermore, information obtained during such "fact finding" visits is used in annual audits. The protection of the restriction on auditing is therefore limited. 12. For practical recommendations concerning the design of a legal frame- work for secured transactions, see Holden and Sahakyan (2004). Chapter 12 Armenia's External Performance and Policy Remedies INTRODUCTION Armenia's macroeconomic growth performance has been impressive over the last half decade.1 So has its foreign trade performance, with exports having emerged recently as the major driver of GDP growth. Yet unemployment continues to stay at double-digit levels and the brain drain has not been reversed, which points to limited employ- ment opportunities across a wide range of skills in Armenia. A World Bank study (2002a) argued that to date Armenia's output growth had not been accompanied by employment generation and poverty reduc- tion, primarily because the business environment was not supportive of enterprise restructuring and new entry for most of the period. From the perspective of 2005, nothing points to any change in this respect. It appears that exports have also had a limited contribution to employ- ment growth. What can be done to increase growth rates beyond the levels already achieved and, simultaneously, reduce unemployment and poverty? Missed opportunities in the export of goods may be linked to four major barriers that prevent Armenia from fully reaping the benefits offered by globalization and that may jeopardize its impressive export growth performance: stalled structural reforms, transport, network- based information technology, and the customs/VAT rebate regime. Weaknesses in these areas have been responsible for some atypical features of Armenia's export performance, namely, the underrepresen- tation of unskilled labor-intensive products in its export basket due in large part to the limited participation in the clothing global value chains and the virtual absence of Armenian firms in a new division of labor spurred by production fragmentation. Network trade has been the most rapidly growing component of world trade over the last two 349 350 Part II: Detailed Analysis decades. Except for diamonds and software, Armenia appears to have missed this opportunity. Armenia's two export success stories--diamonds and software-- corroborate, with an ironic twist, the above assessment. While the two are different, drawing policy conclusions from their success stories would make it possible to circumvent the weaknesses of Armenia's regime in shaping external interaction as well as to soften the external adversities that are due to the country's geography. More important, these stories demonstrate that benefits from reforms that addressed the above weaknesses might be quick to come because they would tap into two great assets--the high quality of human capital and the Armenian diaspora. Several observations emerge from this discussion: Armenia has largely completed the process of implementing first-generation reforms. Prices are not only stable but are also almost fully shaped by market forces. Armenia's exchange and foreign trade regimes are liberal and do not obstruct trade and financial flows. Progress in the implementation of structural reforms has been uneven. For the last three years reforms appear to have stalled. Consequently, there has been no perceptible change in FDI inflows, although they remain some of the highest among CIS economies. While structural reforms take time, some measures that would improve Armenia's attractiveness can be implemented with the stroke of a pen. These include establishing a "white list" of com- panies to receive "flowing treatment" by customs, and a quick VAT rebate scheme. Other recommended measures--better regulatory oversight of telecommunications and unilateral open sky liberalization--may take time to implement, but they should be addressed urgently. Liberalization of trade in services may improve the quality and availability of services through competition, economies of scale, and, last but not least, incentives to policy makers to improve the regulatory environment. The benefits of services liberalization are not limited to the services sectors themselves: they affect all other economic activi- ties. In view of the fact that services contribute on average around 10­20 percent to the production cost of a product and account for all trading costs (transport, trade finance, insurance, communications, and distribution services), the savings from stronger competition from foreign providers can indeed be substantial, as can gains in competi- tiveness in the international markets of both services and goods. Armenia's External Performance and Policy Remedies 351 FOREIGN TRADE AND UNTAPPED OPPORTUNITIES: PERFORMANCE, SPILLOVERS, AND FACTOR ENDOWMENTS Considering the adverse geopolitical conditions peculiar to Armenia and the country's lackluster trade performance through most of the 1990s, the very strong export growth performance after the contraction preceding the Russian financial crisis in August 1998 comes as sur- prise. Contrary to various assessments following the economic stagna- tion in 1998­99 the Armenian economy has displayed surprising vital- ity, driven mainly by its impressive export growth performance. This vitality indicates an important turnaround in Armenia's post-central planning economic performance in three important dimensions. First, the turnaround is based on restructured industrial capaci- ties. The inherited industrial capacities that were incompatible with market disciplines appear to have been mostly dismantled; the emerg- ing industrial structure meets the demanding requirements of inter- national markets, as witnessed by the export growth performance. No other CIS economy recorded such strong growth in 1999­2003. Armenian firms have also outperformed other non-oil CIS competi- tors in both the CIS (Russian) and the EU markets, with their shares in the respective markets increasing significantly in the 1999­2003 period. Their share in total EU imports doubled in 2003 alone. Against the background of falling Russian imports from most CIS countries, the Armenian performance appears particularly impressive, with the value of exports of goods more than doubling between 1999 and 2003. Trade with the CIS appears to be driven no longer by the post-Soviet hysteria in trade patterns, with the emerging trade reflecting a com- parative advantage in these now increasingly competitive markets as these countries make progress in structural reforms. Second, the readjustment in the geographical pattern of trade reflecting the economic weight of regional markets appears to be com- plete. While in 1995 Russia, together with other CIS countries, took 56 percent of Armenia's exports and supplied 49 percent of Arme- nia's imports of goods, these shares fell to 25 percent and 27 percent, respectively, in 1999, and 19 percent and 23 percent, respectively, in 2003. The shift has been largely toward the EU-15,2 whose share in Armenian exports rose from 26 percent in 1995 to 48 percent in 2003 and in imports from 15 percent to 32 percent over the same period. Last but not least, the expansion in exports has not been confined to goods but has also included services, especially if the estimate putting exports of ITC services at around US$100 million rather than US$11 million reported in the balance-of-payments category of "com- puter and information services" is roughly correct. With or without a revised figure for these services, revenues from services increased 352 Part II: Detailed Analysis more than expenditures. With a revised figure, the balance of trade in services swung to a surplus in 2003, and the overall deficit in trade in goods and services was significantly lower. Export Performance in a CIS Perspective Armenia's exports of goods and services--even excluding the estimate of unreported exports of ICT services--in terms of value recorded had the largest growth among the CIS peers in 1999­2003, well above the next best performer, Turkmenistan. Although it might be argued that the growth occurred against a very low base, the fall in trade over the 1990s was not a distinctive feature of Armenia's performance. All CIS countries experienced a huge slump in trade before and immediately after the collapse of the FSU, estimated at around 80 percent between 1990 and 1993 (Michalopoulos and Tarr 1994, p. 5). Although the 1998 financial crisis in Russia punctured the export recovery then under way in all CIS countries (World Bank 2004), exports had contracted earlier in terms of value. In order to account for the differences in the "initial base," as a baseline, Table 12.1 tabulates the index for 2003 using the peak value of exports in a year preceding the contraction (for the respective years, see note (1) to Table 12.1). Except for Azerbai- jan, exports of other CIS countries increased less than in Armenia. Sales to customers in Armenia's two most important markets--the EU and Russia--have driven export growth. These two markets jointly took 60 percent of its total exports in 2003. Armenia was the only country in the CIS that succeeded in readjusting its export offer to the dramatically expanding and shifting Russian import demand toward more sophisticated, higher-quality goods. While suppliers from highly developed countries outperformed Armenian exporters in Russian markets, these were not firms from other CIS countries. Armenia was the only CIS economy whose share in Russia's total imports increased over 1999­2003. Armenia's performance in EU markets has also been superior over the entire 1998­2003 period, except for the contraction in 2001. While, all other CIS non-oil exporters have increased their pres- ence in EU markets relative to other suppliers, except for the Kyrgyz Republic and Uzbekistan, Armenia's gains in import shares were the largest in the 1999­2003 period. Although Armenia's exports as share of GDP also recorded the strongest performance among the CIS economies, its share of exports in GDP was one of the lowest in the CIS. Despite a much larger popu- lation, Ukraine, with a GDP per capita that is only 7 percent larger than that of Armenia, has exports per capita that are 60 percent higher. Only in Georgia, the Russian Federation, and Uzbekistan was this share lower. However, except for Georgia, with a population 57 percent larger (albeit with a GDP only 20 percent higher), both Russia and Uzbekistan Table 12.1 CIS: Total Exports, Export Growth, and Share of Exports in GDP, 1999­2003 Armenia's Index, Total exports Index, Exports as Index, GDP per capita 2003 per capita 2003 percent of GDP 2003 Country 1999 2003 1999=100 1999 2003 1999=100(1) 1999 2003 Note (1)=100 External Armenia 601 916 152 125 338 271 21 37 267 Azerbaijan 550 838 152 154 351 228 28 42 311 Belarus 1,235 1,224 99 653 929 142 53 76 116 Performance Georgia 561 700 125 148 212 144 26 30 106 Kazakhstan 1,141 1,966 172 468 927 198 41 47 172 Kyrgyz Rep. 245 341 139 103 133 129 42 39 88 Moldova 278 443 159 146 236 162 53 53 85 Russian Federation 1,372 3,018 220 593 966 163 43 32 137 and Tajikistan 167 190 114 107 128 119 64 67 103 Ukraine 658 976 148 353 547 155 54 56 137 Policy Uzbekistan 333 378 113 119 137 115 36 36 83 Source: Derived from the World Bank Sima database (http://sima/datasite/sima-web/default.htm). Remedies Note: (1) Data are from 1995 for Tajikistan; 1996 for Armenia, Russia, and Ukraine; and 1997 for the remaining eight CIS countries. These are the years when the value of exports of goods peaked before it subsequently fell in 1997 and 1998. (2) Net exporters of oil and gas are marked in bold. No data were available for Turkmenistan. 353 354 Part II: Detailed Analysis have a much larger domestic economy. Furthermore, exports of goods and services as a percent of Armenia's GDP were also significantly lower than in Moldova (53 percent in 2003), with a smaller GDP per capita, and more developed in Lithuania (56 percent in 2003). Hence, the level of openness of the Armenian economy remains low and its prosperity critically depends on an acceleration in trade growth performance. Without strong export growth well above GDP growth, Armenia will not catch up with more developed economies. Commercial Services and Nondiamond Exports Since 1999 exports have been growing faster than GDP on an annual basis, with the differential sharply increasing in 2002­03 (Figure 12.1). The share of exports of goods and services in GDP rose from 19 percent in 1998 to 37 percent in 2003, while the share of imports fell from 53 percent to 49 percent. Consequently, the import cover- age by exports of goods and services has improved: exports in terms of imports increased from 42 percent in 1999 to 63 percent in 2002 and 64 percent in 2003. Moreover, a strong GDP growth performance considerably lessened an overall potentially negative macroeconomic impact of trade imbalances. In terms of GDP, the trade deficit has been falling rather rapidly, from 38 percent in 1997 to 27 percent in 2000 and 12 percent in 2003. While the annual growth rates of exports of services, as captured in balance of payments statistics, lagged behind the growth of GDP in 2001­03, exports of goods were the driving force behind Arme- nia's GDP growth performance. But international statistics may fail to account for exports of services. In fact, it appears that exports of services were an important factor behind Armenia's impressive exter- nal performance. Trade in Services Growth in Armenia's trade in services is well below the growth of trade in goods. The share of commercial services (excluding govern- ment transactions) in Armenia's total exports of goods and services peaked in 1998 at 34 percent and fell to 25 percent in 2002 and 19 percent in 2003. But the "contraction" or "stagnation" in exports of services is rela- tive and does not suggest that the services sector is not on a sound footing. The relative decline has occurred against a blockbuster per- formance of trade in goods and the rapid growth of the GDP. The value of exports of services increased 53 percent between 1999 and 2003. This would have been an impressive performance had there been a weaker performance in the trade in goods. But over the 1999­ Armenia's External Performance and Policy Remedies 355 Figure 12.1 Dynamics of Exports of Goods and Services and GDP and Change in Exports and Imports as Percent of GDP, 1998­2003 60 60 50 50 40 percent 40 in percent 30 in rate 30 20 GDP in growth 20 10 share annual 0 10 -10 0 1998 1999 2000 2001 2002 2003 year imports as percent of GDP annual rate of GDP growth exports as percent of GDP annual rate of export growth Source: IMF Balance of Payments statistics and World Bank Sima database. 2003 period exports of goods more than tripled. Therefore, the matter deserves a closer examination. Trade in services is defined in GATS (the General Agreement on Trade in Services) as the supply of a service through any of four "modes of supply." Mode 1 is services supplied from one country to another (for example, international telephone calls), officially known as "cross-border supply." Mode 2 consists of consumers from one country making use of a service in another country (for example, tourism), officially known as "consumption abroad." Mode 3 involves a company from one country setting up subsidiaries or branches to provide services in another country (for example, a bank from one country setting up operations in another country), officially known as "commercial presence." Mode 4 involves individuals traveling from their own country to supply services in another country (for example, an actor or a construction worker), officially known as "movement of natural persons." 356 Part II: Detailed Analysis Although specific commitments under GATS are defined in terms of the four standard modes of supply, the trade statistics for the avail- able services are not reported according to a mode of supply and do not allow taking a full account of trade in services. International transactions as defined in the national accounts and balance of pay- ments statistics diverge from the GATS definition of services. Even if some transactions can be captured in balance of payments statistics, the latter suffer from three major deficiencies: they do not distinguish between modes of supply; the definition of a resident and a non- resident differs from the GATS; and the breakdown of services differs from the GATS. The factors of production staying longer than a year in a country are treated as having acquired the status of resident. Fur- thermore, local sales of foreign entities are no longer treated as trans- actions between residents and nonresidents, and as such captured by balance of payments statistics. Transactions falling into such modes of supply as commercial presence and movement of natural persons are not reported in these statistics. Balance of payments statistics do not distinguish between modes of supply (that is, cross-border sup- ply, consumption abroad, commercial presence, or presence of natural persons for less than a year). Furthermore, within WTO definitions, all services except transport and travel are identified as "other commer- cial services" and services exercised by governments are excluded. Hence, as long as a new reference framework for measuring trade in services according to GATS is not in place, proxies derived from the balance of payments statistics and information from other sources are the only way to estimate services trade. The current practice is to use balance of payments statistics categories other than travel and construction as a proxy for cross-border supply (mode 1); "travel" as an indicator of consumption abroad of a service (mode 2); and "com- pensation of employees" together with labor-related statistics from the balance of payments as a proxy for the presence of natural persons (mode 4). The balance of payments statistics do not allow the capture of activities falling under a commercial presence, with the exception of construction services (mode 3). Some estimates of mode 3 can be derived from FAT (foreign affiliate transfers) statistics, if available, or data on FDI stocks together with production statistics such as value added (share of foreign firms in a market) or turnover of foreign firms with their share in respective markets for services. Proxies are what they are--only surrogates for the absence of sta- tistical concepts and methodologies that would allow gauging trade in line with the GATS classification. The major problems stem from the two following weaknesses of balance of payments statistics from the point of view of the task at hand. First, the balance of payments category "travel" tends to overestimate trade in mode 1, as it includes Armenia's External Performance and Policy Remedies 357 consumption of goods. The statistics do not provide information that allows for distinguishing what tourists and business travelers spend on hotels, transportation, services in restaurants, and foods and other goods. Second, some balance of payments components may fall into more than one mode of delivery; that is, it is impossible to assess how a service has been provided. For instance, the transaction between a resident and a nonresident of a country concerning the provision of a computer-related service may take place over the Internet (cross-border, mode 1) or on the site of a resident firm by an expert dispatched by a foreign firm (commercial presence, mode 4). Taking all these factors into consideration, the data presented in Table 12.2, limited to the balance of payments and FDI statistics, seek to capture trade in services according to the mode of supply. Not sur- prisingly, transborder trade in services (mode 1), comprising all bal- ance of payments categories except travel and construction, accounts for the largest share. Construction activities and information on FDI in services are used to assess the scope of trade under mode 3. No data have been available to assess the services trade related to the presence of natural persons (mode 4). Commercial presence (mode 3--that is, the supply of services by foreign-owned firms) is of particular interest, as it potentially has a positive impact on the business climate, provided that a competitive environment is in place. While we do not have data on the foreign firms' share of the total value of services provided by various sectors, the total stock of FDI in services appears to be quite large. On average, FDI inflows in services amounted to 3 percent of GDP over 1998­2003. Their impact on employment might have been considerably higher than the FDI figures alone might imply. For instance, provision of software services--a trademark of Armenia's rapidly developing spe- cialization--is a knowledge-intensive rather than capital-intensive activity, and such low-capital activities have been rapidly expanding in Armenia. Lycos Armenia, a subsidiary of a large multinational soft- ware corporation, is an interesting example (Box 12.1). However, some of these inflows have failed to generate high-qual- ity and low-price services. The primary example is the telecommuni- cations services. This sector accounted for 43 percent of total FDI flows to services over 1998­2003, but the restriction of competition in this sector (as discussed in chapter 10 of this book) has meant expensive and poor quality services. The existence of competitive markets for services combined with the commercial presence of foreign-owned firms is crucial to trade in both goods and services. FDI amounting to US$25 million invested in hotels and restaurants has probably contributed to exports in mode 2--that is, to tourism. FDI peaked over 1998­2000, whereas 358 Part II: Detailed Analysis Table 12.2 Trade in Services by Mode of Supply, 1999­2003 (US$ million) Exports (revenues) Imports (expenditures) Mode of supply 1999 2000 2001 2002 2003 1999 2000 2001 2002 2003 Mode 1: Transborder supply 93.6 91.7 112.2 104.6 115.9 142.9 136.5 146.5 155.2 170.1 Mode 2: Consumption abroad 31.0 37.8 64.9 63.3 71.3 36.7 39.7 39.7 53.9 68.5 Mode 3: Commercial presence Source: Own calculations based on Armenian official Balance of Payments statistics. Armenia's External Performance and Policy Remedies 359 Box 12.1 Lycos Armenia on the Move Lycos Armenia provides four insights into the development of ICT serv- ices in Armenia. First, it shows the ability of Armenians to build upon human skills developed before independence. Lycos Europe (originally founded in May 1997 as a joint venture between Lycos Inc., USA, and Bertelsmann AG, Germany)a, a leading European Internet firm running one of the most popular Internet portals, moved to Armenia in August 2002 with the purchase of the local branch of Brience Inc., a mobile software solution provider company based in San Francisco. Brience Inc. did not establish the company in Armenia but acquired in 2000 a local Armenian company, started in 1986, that was one of first users worldwide of Java technology. Second, the case of Lycos Armenia illustrates strong growth potential and Armenia's strong comparative advantage in ICT-related activities. Lycos Armenia has become the largest developer of software within the Lycos Europe network. The staff of Lycos Armenia grew from 55 to 170 in early 2005, with plans to further increase to 220 employees over 2005. The French media reported that the company intends to lay off 75 employees in its Paris division and has offered 34 of them employment in its expanding subsidiary in Armenia. Third, these are highly paid jobs. According to ICT sources, an expe- rienced software expert makes around US$ 1,000 a month. Last but not least, Lycos Armenia cooperates mainly with other firms of the Lycos network. The commercial presence is designed not to serve local markets but to develop "programmed" products for sale in world markets. Sources: See Lycos' Web site (http://www.lycos-europe.am/pages/history.htm). a. Lycos Europe has branches in all major European capitals, and its total workforce is estimated at 900. The company's single largest shareholder is the U.S. group Terra Lycos, which created one of the world's first Internet search engines in the mid-1990s. the value of exports under mode 2 (mostly tourism) more than dou- bled over 1999­2001. Since exports continued to rise in 2003 despite falling inflows, clearly viable infrastructure for future business and individual travel has been established. The congruence in the time profiles of FDI inflows and exports under mode 2 does not seem to be accidental. Furthermore, investments of US$3.7 million in 2003 in air transportation services or, more precisely, in the modernization of Yerevan Airport (leased to a foreign contractor) have provided an incentive to returning tourists and business persons. The airport also has facilities to handle air cargo efficiently, with a potentially positive impact on trade in goods. 360 Part II: Detailed Analysis The link between commercial presence and trade in goods is both direct and indirect. The direct link is that without logistics-related and other commercial services, no trade on a significant scale can take place. Foreign providers may contribute to services' availability and low prices. The indirect link is the contribution of commercial pres- ence to an improved business climate and an increase in the country's attractiveness to FDI. For instance, the presence of internationally rec- ognized banks and freight integrators may tip the balance in favor of a decision to invest. Their presence has probably contributed to the expansion in exports of ITC services, as it established a safe venue for payments for services. Similarly, the foreign presence exceeding US$10 million over 2001­03 in such areas as R&D and education is likely to have valuable spillover effects that will enhance the competitiveness of knowledge-intensive sectors of the economy. Although Armenia's trade in services does not appear to be strong in contrast to trade in goods, the alleged contraction or stagnation in exports of services is relative and does not suggest that the services sector is not on a sound footing. In fact, it appears that trade in services has been grossly underesti- mated, as statistics fail to capture the lion's share of exports of services provided, for instance, over the Internet. The balance of payments statistics estimate the value of exports of computer and information services (falling in mode 11 of the transborder supply) at US$11 mil- lion in 2003, up from US$1 million in 2002. Other widely quoted esti- mates of ITC exports in Armenia put their value (almost exclusively encompassing Internet solutions, embedded system designs, and tel- ecommunication systems software architectures) at around US$100 million.3 If this is indeed the case, then Armenia has been a huge net exporter of services. Services as a share of total exports were 28 percent rather than the reported 19 percent. In brief, the inclusion of this estimate of ITC exports of services significantly changes the overall picture of Armenia's interaction with the global economy. It suggests a much higher level of openness of the Armenian economy. More important, it also suggests that external interaction has a crucial role in driving Armenia's GDP growth and reducing unemployment. Exports of Goods While diamonds have been major contributors to Armenia's expan- sion of exports of goods, they have not been the only levers of growth. First, an increase in the share of diamonds in total exports until 2003, which grew at an average least-square rate of 44 percent, reflected the impressive growth of all exports, not the result of the stagnat- Armenia's External Performance and Policy Remedies 361 ing exports of nondiamond products. With an annual growth rate of 33 percent over 1999­2003, nondiamond exports also experienced impressive growth. As a result, they caught up with diamond exports in terms of value in 2003, with their share in total exports of almost 50 percent. But in 2004 nondiamond exports continued their impressive growth, while diamond exports contracted (Figure 12.2). Second, the increase in concentration was not merely the result of the growth in diamond exports but also of the deepening export specialization and less marginal presence in world markets. There has been little change in the largest export sectors, as the correlation between the shares of exports of 25 four-digit SITC product groups in 1999 and 2003 is very high at 98 percent. However, the top nine SITC Rev 2 four-digit sectors (excluding diamonds) in exports in 2003 expe- rienced stronger growth than diamonds. Their average annual growth rate was 39 percent in 1999­2003, and the share of diamonds in the top ten sectors fell from 63 percent in 1999 to 60 percent in 2003. What are the nondiamond products that Armenia exports? By far the single most important group is spirits and liqueurs (SITC 1124), whose Figure 12.2 Exports of Diamonds and Other Products, 1997­2004 600 exports of other goods 500 exports of diamonds 400 million 300 US$ 200 100 0 1997 1999 2000 2002 2003 2004 year Source: Derived from data reported by Armenia to the UN COMTRADE database. 362 Part II: Detailed Analysis foreign sales quadrupled between 1999 and 2003 and whose share in total exports of goods rose from 5 percent to 13 percent over this period. Other products include nonferrous base metals (SITC 28882), jewelry (8973), ores and concentrates of nonferrous metals (2879), ferro-alloys (6716), metal-cutting machine tools (7361), aluminum alloys (6842), copper alloys (6821), and copper ores (2871). Taken together, these products contributed 30 percent to total exports in 2003. Armenia appears also to have been regaining a competitive advan- tage in other products in which it used to specialize when it was part of the former Soviet Union. These products include ores and concen- trates of nonferrous metals, liqueurs, tungsten, molybdenum, tanta- lum, magnum, copper, and ferro-alloys. However, except for liqueurs and ores and concentrates of nonferrous metals, the EU has replaced the former Soviet republics as the major export market. In 2003 it took 100 percent of all exports of tungsten, molybdenum, tantalum, and magnum (SITC 6891), 100 percent of exports of ferro-alloys, and 95 percent of exports of copper alloys. Nevertheless, a bad sign is that manufactures other than diamonds were conspicuously absent in Armenia's exports. Although the share of manufactures excluding chemicals increased in Armenia's total exports from 39 percent in 2000 to 60 percent in 2003, this was mainly due to the increase in exports of diamonds. Other manufactured exports fell in terms of value from US$59 million in 2000 to US$44 million in 2002 and increased to US$54 million in 2003. Their share in total exports fell from 29 percent in 2000 to 11 percent in 2002 and 10 percent in 2003. While the sluggish growth of manufactured exports other than dia- monds in the presence of Armenia's potential comparative advantage should raise policy concerns, there is nothing inherently wrong with a large concentration of exports. High concentration is the result of the small size of the national economy and its narrow industrial base. There is one important stipulation, however. If the bulk of export reve- nues comes from a commodity subject to weather conditions and high volatility in international markets, then a national economy may be vulnerable to external shocks. But this is not the case with diamonds and Armenia. Diamonds seem to be less vulnerable than other "sin- gle crops," including textiles and clothing, to international supply or demand volatility for two reasons. First, Armenian firms are firmly entrenched in a global diamond value chain through international investors, albeit not completely invul- nerable. Armenian diamond-polishing factories are parts of both the Antwerp-centered link and the Israeli-based Lev Leviev Group. There is an important caveat, however: some activities depend on supplies from Russia. A contraction of almost 75 percent in the Russian supply of Armenia's External Performance and Policy Remedies 363 raw diamonds in 2004 has affected Armenia's diamond export perform- ance. Thus in 2004, the value of Armenia's diamond exports contracted 33 percent. Both total and EU-destined exports, however, increased in terms of value, but exports to Israel, one of the major recipients of dia- monds cut in Armenia, significantly contracted (Table 12.3). Second, the diamond market remains fairly controlled, with prices displaying remarkable stability. Although previous performance is not necessarily an indication of future performance, a sudden fall in prices and demand is rather unlikely. As The Economist (2004) notes, "the cartel isn't forever." However, it adds that no major player would seek to bring the prices down. Yet spillovers from diamond production appear to be relatively lim- ited, as value added locally is limited to employing the local labor force and using energy and other services. There are no other inputs into diamond production, which is mainly a gem-cutting operation. Thus, the high value of diamond exports, accounting for half of total exports, overstates their overall impact on economic welfare. On the other hand, one may argue that reliance of exports on just one sector of the economy does not set Armenia aside from many other transition economies. For instance, the share of the 10 best- performing sectors in total exports exceeded two-thirds in 2003 in at least five transition economies--Albania, Azerbaijan, Kazakhstan, the Kyrgyz Republic, and Moldova. A single sector was responsible for around half (or more) of the total exports of Azerbaijan, Kazakhstan, and the Kyrgyz Republic (Table 12.4). There were different sectors involved--gold mining in the Kyrgyz Republic, oil in Azerbaijan and Kazakhstan, and textiles and clothing in Albania, Bulgaria, Macedo- nia and Moldova, albeit the concentration of exports in the last was significantly lower. Both oil and clothing, however, have different characteristics from diamonds, as they usually involve higher value left in a country (oil), higher value added in a country, or are produced locally. The cloth- ing sector is more labor intensive than diamond cutting and has the potential to generate spillover effects provided the economic regime is friendly toward private business activity. Revenues from oil, if not mismanaged by central authorities, can contribute to the development of a country's infrastructure, thereby facilitating trade. These observations should not diminish Armenia's success in find- ing an attractive niche in global markets. But they add urgency to adopting measures that would expand Armenia's export offer. Other exports are already emerging, indicating the growing specialization and falling marginalization of Armenian suppliers in international markets. It is hoped that diamonds will continue to shine, providing employment and reducing poverty. 364 Part II: Detailed Table 12.3 Recipients of Armenia's Exports, 2000­04 (US$ million) Analysis Index, 2004 Share of exports 2000 2001 2002 2003 2004 2000 = 100 European Union 247 196 245 351 401 162 Israel 2 20 80 123 100 5,005 Total 300 342 347 566 591 465 Total annual change n.a. 14% 2% 63% 4% 197 Memorandum share in world exports 0.005% 0.006% 0.005% 0.008% 0.007% 138 EU-15 imports (including intra-EU trade) 0.005% 0.004% 0.005% 0.008% 0.006% 139 Source: Own calculations based on data from the IMF Direction of Trade database (http://www.imf.org/external/data.htm#data). n.a. not applicable Armenia's External Performance and Policy Remedies 365 Table 12.4 Sector Concentration of Exports in Selected Transition Economies, 2003 (percent) Share of Share of the largest Share of top ten item in a oil and SITC Rev 2 SITC Rev 2 Share of oil-related Country 4-digit sectors 4-digit sector clothing products Albania 69 26 34 n.a. Armenia 83 50 n.a. n.a. Azerbaijan 91 70 n.a. 85 Bulgaria 29 6 20 n.a. Kazakhstan 78 45 n.a. 58 Kyrgyz Republic 74 45 n.a. n.a. Macedonia 45 7 30 n.a. Moldova 53 28 15 n.a. Source: Own calculations based on the UN COMTRADE database (http://unstats.un.org/unsd/comtrade/default.aspx). Note: n.a. stands for not applicable; that is, the share in total exports is below 5 percent. Low Participation of Unskilled Labor in Exports Armenia's factor endowments suggest specialization in natural resources and labor-intensive products. Armenia has significant deposits of copper and zinc and traditionally has been a producer of industrial minerals. It is amply endowed in clay, construction stones, and semiprecious gems, with small amounts of mined gold as well. Armenia also has climate conditions that favor agriculture and is noted for the quality of its fruits and grapes. Last but not least, it has a large pool of a relatively low-cost labor, both highly skilled and unskilled. Hence, products that are intensive in the use of unskilled labor and in the use of natural resources should dominate its export basket. On the import side, Armenia should be a large importer of capital and skilled labor-intensive products together with fuels. The factor intensities of Armenia's exports of goods, however, do not fully corroborate these expectations. Armenia is a net exporter of skilled labor-intensive products and a net importer of all other products, including unskilled labor-intensive and natural resource- intensive products (Table 12.5). But in line with its endowments, natural resource-intensive products accounted in 2003 for more than half of Armenia's total exports. The evolution and the current factor content of Armenia's exports point to two conflicting developments: the emergence of market disci- pline and the persistence of domestic barriers to the efficient allocation 366 Part II: Detailed Analysis Table 12.5 Factor Intensities of Armenian Trade in 1997, 2000, and 2003 Export value Export share (in US$ million) (in percent) Factor intensity 1997 2000 2003 1997 2000 2003 Natural resources 98 34 284 45.8 17.4 52.6 Unskilled labor 12 6 5 5.7 3.3 0.9 Capital 35 34 17 16.3 17.4 3.1 Skilled labor 63 111 214 29.3 57.1 39.7 Total 213 195 540 100 100 100 Exports in percent Import share of imports (in percent) Factor intensity 1997 2000 2003 1997 2000 2003 Natural resources 24.2 7.8 39.2 53.6 52.3 62.3 Unskilled labor 34.0 13.9 6.3 4.7 5.5 6.5 Capital 25.1 18.9 8.3 18.4 21.8 17.4 Skilled labor 94.1 120.6 133.4 8.8 11.1 13.8 Total 28.3 23.5 46.3 100.0 100.0 100.0 Source: Own calculations based on Armenia's national data as reported to the UN COMTRADE database (http://unstats.un.org/unsd/comtrade/default.aspx). Note: Factors are taken from SITC-4 Rev. 1. of resources in the national economy. The marginalization of exports of capital-intensive products points to an ongoing industrial restruc- turing in line with Armenia's endowments, illustrated by the disap- pearance of industrial capacities, inherited from Soviet-style industri- alization, that were noncompetitive without government subsidies. This readjustment occurred in 2000­03, with the share of capital-inten- sive products falling from 17 percent to 3 percent (Table 12.5). Their exports in terms of value fell 50 percent between 2000 and 2003. While the fall in capital-intensive exports provides evidence of the implementation of market-oriented reform measures, a very low share of unskilled labor-intensive products in Armenia's exports points to unfinished reforms. While among the CEEC-10 and SEE-5 economies the lowest share of unskilled labor-intensive products was 10 percent (Hungary), in 2003 this share in Armenia's exports--as well as in the exports of four other CIS economies (Azerbaijan, Georgia, Russia, and Kazakhstan)--was a mere 1 percent. The contraction in the share of unskilled labor-intensive products was not the result of faster growth of other exports, as the value of exports of unskilled labor-intensive products was lower in 2003 than in 2000. Similarly and more gener- Armenia's External Performance and Policy Remedies 367 ally, stagnation in the exports of manufactures other than diamonds in 2000­03 points to the inability of tapping Armenia's major asset--the availability of a cheap and disciplined labor force. Hence, although open economies based on competitive markets have a unique trade pattern reflecting their endowment, this is yet to become the case with Armenia. The discord between Armenia's exports and its endowments in production factors points to either a deficiency in the mechanisms allocating capital and labor in the economy or to adverse external factors beyond government control. The former has to do with the low quality of the business environment, whereas the latter may include protectionist conditions in market access or high transportation costs. Access to most markets, excluding Azerbaijan and Turkey, is not a barrier, as it is either preferential (as with the CIS, EU-25 and other highly developed OECD economies) or based on Armenia's most favored nation status. But transportation is clearly a barrier in most land-locked countries. High transportation costs favor exports of those products with high value relative to weight and dis- courage sales of the bulky low-cost products of light industry, which usually involve unskilled labor-intensive activities. Whatever the explanation, the low share of unskilled labor-intensive products in Armenia's exports while unemployment remains at double- digit levels points to Armenia's inability to tap the opportunities offered by global markets. Furthermore, even a quick examination of Armenia's exports sug- gests that to date they have contributed little to reducing unemploy- ment and poverty. Unemployment has been locked for more than a decade at high double-digit levels (around 30 percent), two-thirds of the formally employed continue to be in the low value-added sectors of the economy (agriculture and trade), and one-third of Armenians were below the poverty line in 2003. In 2003, McKinsey warned that the Armenian GDP growth pat- tern, relying on a narrow base and generating limited employment effects, would encounter difficulties by 2005­06 (McKinsey & Com- pany 2003). Exports can help to reverse the current growth pattern provided that reform measures remove barriers to business activity. Once the barriers are removed, investors would take advantage of Armenia's assets--an educated labor force, the low labor cost, and the presence of the diaspora--and would turn the potential comparative advantage into a revealed one. Untapped Opportunities in Global Markets The idiosyncrasy of Armenia's export basket in terms of factor inten- sities stems, on the one hand, from its ability to tap the benefits of becoming part of a global value chain for diamonds and, on the other 368 Part II: Detailed Analysis hand, from its inability to take advantage of opportunities created by the emerging forms of the global division of labor. The outsourcing of services and production activities has been the driving force of world trade. Thanks to technological progress and the information revolution, it has become possible to divide the industry's value chain, including services, into smaller functions that can be contracted out to independent suppliers. This fragmentation of production has offered unique opportunities for producers in small countries to move from servicing their limited local markets to supplying large multinational firms and, indirectly, their customers throughout the world. The major sectors in which this process has been taking place are textiles and clothing, footwear, furniture, automobiles, computers, semiconduc- tors, and heavy machinery. Automobiles and ICT represent the most dynamic networks. The sequence observed in successful European transition economies was the movement from textiles and clothing, footwear, and furniture to automobiles, heavy machinery, and ICT. Armenia has successfully entered the diamond global value chain but has yet to participate in other value chains and networks, including first-stage simple networks. Textiles and clothing have almost disappeared in Armenian exports, as the value of their exports fell from US$13 million in 1999 to US$1.7 million in 2003. As for another easily accessible network, Armenia has never been an exporter of furniture and its parts. The situation has not changed, as Armenian firms have not plugged into the global supply chains of multinational retailers of furniture. Exports of furniture and its parts have been practically nonexistent: their value was well below US$100,000 in 2003. Nor have Armenian firms become part of the second- stage, more sophisticated, networks of production and distribution. Given the country's location and lack of tradition, it is not surpris- ing that no Armenian firm has become a supplier to major producers of automobiles. However, Armenia's absence from the global supply chains of ICT firms comes as a surprise. Despite outstanding success in the development of the ICT sector, nothing indicates the participation of Armenian firms in e-networks (such as electronics, including office equipment and telecommunications). ICT firms have not been able to take full advantage of huge opportunities for outsourcing created by recent changes in the ICT sector worldwide. With the disappearance of the global "one-stop-shop" industry structures in the early 1990s,4 opportunities have emerged for new entrants. Armenian firms have successfully entered the software and imaging technology niches. But they have failed to enter other stages of the production and delivery processes--in particular, as providers of front-end customer contact/ support services or suppliers of components (for example, metal, plas- tics, and electronic components).5 Armenia's External Performance and Policy Remedies 369 In all, the manufacturing sector remains underrepresented in Arme- nia's export offer. It has clearly underperformed both domestically and internationally, as construction and other services have driven GDP growth and diamonds have driven export growth. This repre- sents an important missed opportunity. Owing to its ample endow- ment in skilled labor, Armenia's strongest comparative advantage lies in the manufacturing sector. FDI: Missed Opportunities and Factor Intensities of Armenian Exports The discord in Armenia's exports in terms of factor intensities reminds one of similar discords revealed in the exports of many economies dur- ing the initial stages of the transition from central planning.6 Former Czechoslovakia, Hungary, and Poland experienced this in the early 1990s. FDI inflows have subsequently been responsible for closing the gap between the endowments and the factor content of exports. Trade in network products in particular has led to the increase in shares of skilled labor- and capital-intensive products in their exports. FDI has yet to accomplish the same task in Armenia for two reasons. First, FDI inflows have been relatively low. Armenia has performed very well in terms of attracting FDI by CIS standards, excluding oil- rich countries (Azerbaijan, Kazakhstan, and Russia).7 But these have not been very demanding benchmarks. It should be borne in mind that its FDI inflows per capita were significantly lower than the simple average of FDI per capita in the CEEC-8 (Table 12.6). Second, although FDI as a percent of Armenia's GDP was quite significant (only 0.3 percentage points below the simple average for CEEC-8 economies), this would not change the picture significantly in terms of its impact on trade in part for at least three reasons. First, FDI inflows into Armenia are of relatively recent vintage. Until the sale of 90 percent of ArmenTel to OTE of Greece for US$142.5 million in late 1997,8 the total FDI inflows into Armenia in 1990­97 were less than US$50 million.9 Second, almost one-fifth of the total FDI inflows over 1995­2003 went into telecommunications with rather disappointing results, as current penetration rates for fixed and mobile services in Armenia remain significantly below the regional benchmarks, result- ing in a rather marginal improvement in the quality of these services. Third, other infrastructure-related sectors (such as water, electricity, and aviation have accounted for a sizable share of total FDI inflows (Table 12.6). While these inflows, as well as investment in hotels, restaurants, and so forth, are important for creating an environment enabling business activity, it usually takes time before they have a positive impact on such activities. 370 Part II: Detailed Table 12.6 FDI Inflows Compared, 1995­2003 Average Countries 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2000­03 Analysis FDI per capita Armenia 2 5 5 58 42 39 24 47 49 74 40 CIS-11 (simple average) 7 16 23 36 28 22 30 38 57 36 CIS-8 (excluding oil-rich) 5 7 14 24 16 15 15 20 22 18 CEEC-8 (simple average) 123 100 102 123 100 102 171 220 239 183 FDI as percent of GDP Armenia 0.5 1.1 1.1 12.3 7.0 6.3 3.6 6.0 5.5 5.5 5.3 CIS-11 (simple average) 1.1 3.0 3.7 5.7 4.1 3.3 3.6 4.5 6.8 6.8 4.5 CIS-8 (excluding oil-rich) 1.1 1.4 2.4 3.5 2.6 3.0 3.0 2.8 3.2 3.2 3.0 CEEC-8 (simple average) 3.7 3.1 3.2 5.5 4.8 5.8 5.6 7.4 5.0 5.0 5.8 Source: NSS and EBRD Annual Transition Reports (various issues) for other countries. Armenia's External Performance and Policy Remedies 371 Indeed, there are reasons to believe that their impact on both exports and patterns of growth will be felt soon. It should be considered, first, that foreign inflows into Armenia have not consisted merely of FDI. In fact, Armenia has been the largest recipient of foreign aid per capita among CIS countries and, in addition, foreign remittances have amounted to around one-fifth of GDP on average for the last six years. Kuznetsov and Sabel (2004) rightly observe that remittances, private transfers from diaspora Armenians, and FDI inflows should provide a springboard for business development and entrepreneur- ship, provided that the institutional environment becomes more busi- ness friendly. Second, the high-quality FDI inflows--that is, those attracted by prospects of returns not based on special rents negotiated with the government (as was the case with the privatization of telecommunica- tions)--account for the growing share of FDI inward stock in Arme- nia. Total FDI in 2002­04 accounted for almost half (49 percent) of the total FDI inflows to Armenia since independence. This coincides with significant progress being made in implementing structural reforms, which usually augurs well for attracting high quality investors. So does the fact that 53 percent of total FDI over 2002­04 was not related to large-scale privatization (Figure 12.3). The composition of FDI inflows suggests that efficiency gains may be accruing in certain sectors to the benefit of potential exports. First, while the proportion of cumulative FDI inflows over 1995­2003 going to the industrial sector (56 percent) is in line with similar shares in the CEEC-8 economies, most of these investments have been concen- trated in two industries--electricity and beverage production. These two industries took almost two-thirds of cumulative inflows going to industries. Mining absorbed another 15 percent of cumulative FDI flows to the industrial sector. Exports of beverages (brandy) soared, confirming the positive impact of FDI on exports. Their value more than doubled in 2000 alone, with their share in total exports rising from 5 percent in 1999 to 8 percent in 2000 and 13 percent in 2003. Similarly, one suspects that the revival of exports of ores was linked to FDI. These exports collapsed in 1999 and reached the 1997 level only in 2002. Ores and other mining were among the largest recipients of FDI in 1998­99, taking US$50 million, or 27 percent, of all industrial FDI at that time. In addition, the successful privatization of the Zange- zur copper-molybdenum plant in Kajaran to Germany's Chronimet will dramatically increase the degree of processing embodied in the exports of molybdenum (Grigorian 2005). Second, there were also some investments in the manufacturing sector that may increase export potential. For instance, the medical, optical and measurement equipment area and the communications 372 Part II: Detailed Analysis Figure 12.3 Privatization and FDI Inflows, 1998­2004 200 100 80 150 percent in 60 millions 100 FDI 40 US$ total in 50 20 share 0 0 1998 1999 2000 2001 2002 2003 2004 year other 46% 48% 66% 64% 64% 57% 69% privatization 54% 53% 34% 36% 36% 43% 31% other 107 64 83 49 95 90 163 privatization 125 71 42 27 54 67 72 Source: Derived from NSS official data as reported in Grigorian (2005). equipment area, both regarded as knowledge-intensive sectors, have attracted significant FDI inflows. In 2003, US$57 million worth of FDI went to medical, optical and measurement equipment, which received US$9 million in 2002­03 (Table 12.7). Furthermore, the almost US$6 million in FDI that went to the furniture sector over 2001­02 may be a harbinger of participation in global furniture networks. But this remains to be seen. In all, although there appears to be a strong correspondence between sectors that received significant inflows of FDI and their export per- formance, the problem remains that FDI inflows did not rush to the labor-intensive manufacturing sectors. By the same token, Armenia has not become part of the sophisticated global production and dis- tribution networks. More significantly, except for diamonds, Armenia did not become part of value chains such as clothing and furniture, where outsourcing does not necessarily require significant inflows of FDI and know-how. Table 12.7 Receipt of FDI Inflows by Sector, 1998­2003 Armenia's US$ million Sector 1998 1999 2000 2001 2002 2003 Total FDI 2003 Share in % Agriculture 0.0 0.0 0.0 0.2 3.4 0.0 3.6 0.4 External Industry 98.9 94.7 56.5 35.3 86.0 102.2 473.6 55.5 Electricity, gas, hot water, and steam production and distribution 43.3 42.0 42.0 24.9 42.4 3.8 198.3 41.9 Performance Food and beverage production 13.2 37.7 7.8 3.9 13.6 12.9 89.0 18.8 Medical, optical, and measurement equipment and watch production 0.0 0.0 0.0 0.9 0.5 57.0 58.5 12.4 Ore (metal-bearing) mining and other subsectors of mining 29.0 10.5 3.4 0.7 3.3 12.2 59.2 12.5 and Chemical industry 2.6 0.2 0.2 0.7 10.8 0.3 14.8 3.1 Policy Metallurgy 0.6 0.5 0.4 2.2 3.0 6.4 13.0 2.7 Radio, broadcasting, and communications equipment production 0.0 0.0 0.0 0.0 2.6 6.0 8.5 1.8 Remedies Services 133.5 35.6 63.5 40.4 51.5 51.3 375.8 44.1 Source: NSS. 373 374 Part II: Detailed Analysis Conclusion: The Cost of Missed Opportunities The cost of missed opportunities seems to be significant. First, where textiles and clothing are concerned, the activity is highly labor-inten- sive, creating employment opportunities for a low-skilled and cheap, usually female, labor force. The welfare cost of the collapse of these exports is very significant. Second, the incorporation of local producers into other produc- tion and marketing networks as well as the supply of ICT consumer services could extend significant benefits to Armenia. This arrange- ment usually brings new technologies and managerial know-how as well as direct access to larger markets and thus brings in the benefits of economies of scale. Networks boost exports without making local firms incur marketing expenses and provide for greater stability in earnings, thanks to the global reach of a "parent" company. The frag- mentation of production eliminates the need to gain competency in all stages and aspects of production and allows a small country to focus on a subset of activities. At the same time, production sharing can broaden the range of final products whose components are produced in the small country and thus protect the country from a demand shock to a particular good. Third, with a stronger export performance, the growth of the Armenian economy could have been potentially higher and more employment-oriented. While exports have recorded a healthy growth since 1999, export growth continues to be concentrated mainly in sec- tors with limited employment opportunities. The expansion of manu- factured exports, especially those associated with global value chains and production and distribution networks, would go a long way toward addressing the challenge of stronger export performance. BUSINESS ENVIRONMENT AND POLICIES OBSTRUCTING EXTERNAL BUSINESS INTERACTION The discord between Armenia's exports and its endowments in pro- duction factors points to either a deficiency in the mechanisms of allo- cation, which discourages domestic and foreign investment, or adverse external factors beyond government control. Governments have lit- tle control over geography, although they can do a lot to lighten the negative impact of remote locations from major world markets. As a beginning, they may address weaknesses in the business environment, improve the contestability of domestic markets, and introduce meas- ures to bring down the cost of entry, exit, and conduct of business activity in both internal and external dimensions. In fact, the experi- Armenia's External Performance and Policy Remedies 375 ence of successful transition economies suggests that two elements have to be in place: the successful implementation of first-generation reforms (such as liberalization of prices, foreign trade, and exchange regimes), and a consistent movement toward a rule-based institutional regime with the capacity of enforcement. The former is relatively easy to implement, in the absence of political opposition, whereas the latter requires the state's advanced institutional capacity. The reasons for Armenia's failure to tap opportunities have nothing to do with first-generation reforms, as Armenia completed the first- generation reform project almost a decade ago. By 1995, prices for most if not all tradables were fully liberalized, small-scale privatiza- tion was completed, the domestic currency had become convertible, and foreign trade and exchange rate regimes were fully liberalized. Inflation, after a period of hyperinflation following the dissolution of the former Soviet Union, appears to be firmly under control. As a result of a tight monetary policy and the appreciation of the national currency, inflation was at 2 percent in 2004. According to international assessments, Armenia has caught up with highly developed econo- mies in first-generation policy areas. However, the benefits of first-generation reforms cannot be fully taken advantage of unless they are accompanied by steady progress in implementing institutional, second-generation reforms. Instead, the unfinished agenda of much more demanding and complex structural reforms together with trade facilitation measures appears to be a bar- rier to a more extensive involvement of Armenian firms in the finer global division of labor based on in-time production, inventory man- agement, and complex Web-based communication links. The question to which we shall now turn is about the progress achieved in struc- tural reforms and their impact on FDI inflows. Unfinished Agenda of Second-Generation Reforms While progress in first-generation reforms is relatively straightfor- ward to assess (for example, by inflation, or convertibility of domestic currency for current account transactions), this is not the case with second-generation reforms that cover, as a rule, activities and policy areas in which progress can be gauged only indirectly. Moreover, these reforms usually tend to be highly intertwined. Progress in one area can be "neutered" by the lack of progress in another. For instance, large-scale privatization may be completed. Yet its benefits may not materialize if privatized companies are subsidized by state-owned banks, as was the case in the Czech Republic in the mid-1990s, or the competition policy is unable to contain the banks' abuse of a domi- nant position. The same applies to various indicators that measure the 376 Part II: Detailed Analysis quality of the business climate. Huge minimum capital requirements or instabilities in tax policies, for example, may offset the simplicity of procedures for business entry. How does Armenia compare with other transition economies in terms of progress achieved in economic reforms? Since Armenia's performance has been superior to that of other CIS economies, we use two other, more demanding, comparators: new Central Euro- pean EU members (CEEC-8), and Lithuania (like Armenia, a former Soviet republic). In order to answer this question, we construct an index of progress in second-generation reforms. The index is derived from EBRD transition indicators of policies that achieved progress in different areas directly related to restructuring. These areas include government and enterprise restructuring, large-scale privatization, banking reform and interest rate liberalization, security markets and nonbank financial institutions, and competition policy. The average of scores in each of these areas--normalized for a minimum score of 1 to equal zero and a maximum score of 4.5 to equal 100--is the indicator of progress in structural reforms. For countries that achieved institutional maturity in terms of establishing institutions supporting competitive markets, the indicator is equal to 100, whereas for those that have not started the process, it equals zero. The indicator of progress in structural reforms also appears to pro- vide a good measure of the quality of governance and the incidence of corruption. The value of the correlation coefficient for the structural progress indicator for 27 transition economies of Europe and Central Asia and corruption perception indices compiled by Transparency International was 73 percent in 2004. The positive correlation is higher for governance indicators (including average of political stability, gov- ernment effectiveness, and regulatory quality),10 as compiled by the World Bank. For the 1998­2002 averages, its value was 93 percent. In other words, countries with higher values of the structural progress indicator are better governed and are less corrupt. The results presented in Figure 12.4 clearly show that moving to the institutional environment of highly developed countries takes time. The median value of a structural aggregate reform index for new EU CEEC-8 member states stood in 2004 at 66 percent of the level of mature market economies. While it is almost two-thirds percent above the level of institutional maturity achieved by Armenia (40 percent), CEEC-8 countries have a long way to go to catch up with highly developed countries. Armenia also has long way to go. The greatest "distance" separat- ing it from institutions in highly developed countries is in competi- tion and security markets and nonbank financial institutions--both at 29 percent of the "normal market-economy" level (Table 12.8). Other Armenia's External Performance and Policy Remedies 377 Figure 12.4 Progress in Structural Reforms: Values of SRI, 1994­2004 (in percent) 70 60 50 40 percent 30 20 10 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 year Lithuania CEEC-8 median Armenia SEE-4 Note: The aggregate is the average of scores ranging between 1 (no liberalization) and 4.5 (liberalization at the levels of highly developed market economies) for government and enterprise restructuring, competition policy, banking reform and interest rate liberalization, security markets and nonbank financial institutions, and large-scale privatization. It has been normalized with 1=0 and 4.5=100. SEE- 4 includes Albania, Bosnia and Herzegovina, Macedonia, and Serbia and Montenegro. Source: Own calculations based on data from EBRD Annual Transition Reports. institutional areas where the value of the index is below its average for second-generation reforms are government an