D I R E C T I O N S I N D E V E L O P M E N T 47812 Finance Adequacy of Retirement Income after Pension Reforms in Central, Eastern, and Southern Europe Eight Country Studies Robert Holzmann and Ufuk Guven Adequacy of Retirement Income after Pension Reforms in Central, Eastern, and Southern Europe Adequacy of Retirement Income after Pension Reforms in Central, Eastern, and Southern Europe © 2009 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 12 11 10 09 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The bound- aries, colors, denominations, and other information shown on any map in this work do not imply any judgement 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 publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work 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; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN: 978-0-8213-7781-9 eISBN: 978-0-8213-7780-2 DOI: 10.1596/978-0-8213-7781-9 Library of Congress Cataloging-in-Publication Data Holzmann, Robert. Adequacy of retirement income after pension reforms in Central, Eastern, and Southern Europe: eight country studies/Robert Holzmann and Ufuk Guven. p. cm. ISBN 978-0-8213-7781-9 (alk. paper) -- ISBN 978-0-8213-7780-2 1. Retirement income--Government policy--Europe, Eastern. 2. Retirement income-- Government policy--Europe, Central. 3. Pensions--Government policy--Europe, Eastern. 4. Pensions--Government policy--Europe, Central. I. Guven, Ufuk, 1972- II. Title. HD7164.7.H647 2009 331.25'2094--dc22 2008041837 Cover photos: Couple in Czech Republic/Corbis; Couple in Romania/Corbis; Couple by Water/Getty. Cover design: Naylor Design, Washington, DC Contents Preface xv Acknowledgments xvii Abbreviations xix Chapter 1 Introduction, Summary, and Policy Conclusions 1 Motivation for Reform and Policy Trends 3 Characteristics of Reformed Pension Systems 10 Assessment of the Performance of Pension Systems 37 Conclusions 52 Notes 56 Bibliography 58 Chapter 2 Bulgaria 61 Motivation for Reform 62 Characteristics of Bulgaria's Pension System 64 Assessment of the Performance of Bulgaria's Pension System 73 Conclusions 84 Notes 85 Bibliography 87 v vi Contents Chapter 3 Croatia 89 Motivation for Reform 90 Characteristics of Croatia's Pension System 91 Assessment of the Performance of Croatia's Pension System 100 Conclusions 110 Notes 111 Bibliography 113 Chapter 4 The Czech Republic 115 Motivation for Reform 116 Characteristics of the Czech Republic's Pension System 117 Assessment of the Performance of the Czech Pension System 127 Conclusions 141 Notes 143 Bibliography 145 Chapter 5 Hungary 147 Motivation for Reform 148 Characteristics of Hungary's Pension System 150 Assessment of the Performance of Hungary's Pension System 162 Conclusions 173 Notes 175 Bibliography 177 Chapter 6 Poland 181 Motivation for Reform 182 Characteristics of Poland's Pension System 184 Assessment of the Performance of Poland's Pension System 195 Conclusions 204 Notes 206 Bibliography 207 Chapter 7 Romania 211 Motivation for Reform 212 Characteristics of Romania's Pension System 213 Contents vii Assessment of the Performance of Romania's Pension System 223 Conclusions 234 Notes 235 Bibliography 237 Chapter 8 The Slovak Republic 239 Motivation for Reform 240 Characteristics of the Slovak Republic's Pension System 242 Assessment of the Performance of the Slovak Pension System 252 Conclusions 262 Notes 264 Bibliography 265 Chapter 9 Slovenia 267 Motivation for Reform 268 Characteristics of Slovenia's Pension System 269 Assessment of the Performance of Slovenia's Pension System 278 Conclusions 288 Notes 289 Bibliography 291 Index 295 Box 1.1 Taxation of Retirement Savings 15 Figures 1.1 Projected Pension System Fiscal Balances before Reform in Eight CESE Countries 4 1.2 Old-Age Dependency Ratios, 2000­50, by World Region 5 1.3 Gross Replacement Rates for Male Full-Career Workers in Eight CESE Countries 41 1.4 Net Replacement Rates for Male Full-Career Workers in Eight CESE Countries 43 viii Contents 1.5 Net Replacement Rates for Male Partial-Career Workers in Eight CESE Countries 45 1.6 Impact of Indexation on Income Replacement (Active Earnings Units) in Eight CESE Countries 47 1.7 Projected Pension System Fiscal Balances after Reform in Eight CESE Countries 50 2.1 Projected Fiscal Balance of Bulgaria's Public Pension System before Reform, 2001­50 63 2.2 Projected Old-Age Dependency Ratio in Bulgaria, 2005­50 64 2.3 Sources of Gross Replacement Rates in Bulgaria, by Income Level 78 2.4 Sources of Net Replacement Rates in Bulgaria, by Income Level 79 2.5 Net Replacement Rates for Male Full-Career Workers in Bulgaria, Europe and Central Asia, and the World, by Income Level 80 2.6 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Bulgaria, by Career Type and Exit Age 81 2.7 Projected Fiscal Balance of Bulgaria's Public Pension System after Reform, 2001­50 82 2.8 Net Replacement Rates for Men in Bulgaria before and after Benefit Adjustment 84 3.1 Projected Old-Age Dependency Ratio in Croatia, 2006­50 91 3.2 Sources of Gross Replacement Rates in Croatia, by Income Level 105 3.3 Sources of Net Replacement Rates in Croatia, by Income Level 106 3.4 Net Replacement Rates for Male Full-Career Workers in 2040 in Croatia, Europe and Central Asia, and the World 107 3.5 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Croatia, by Career Type and Exit Age 108 3.6 Projected Fiscal Balance of Croatia's Public Pension System after Reform, 2005­40 110 4.1 Sources of Gross Replacement Rates in the Czech Republic, by Income Level 132 Contents ix 4.2 Sources of Net Replacement Rates in the Czech Republic, by Income Level 133 4.3 Net Replacement Rates for Male Full-Career Workers in the Czech Republic, Europe and Central Asia, and the World, by Income Level 134 4.4 Net Replacement Rates for Male Low-Income Partial-Career Workers in the Czech Republic, by Career Type and Exit Age 135 4.5 Net Replacement Rates for Male Middle-Income Partial-Career Workers in the Czech Republic, by Career Type and Exit Age 136 4.6 Net Replacement Rates for Male High-Income Partial-Career Workers in the Czech Republic, by Career Type and Exit Age 137 4.7 Projected Fiscal Balance of the Public Pension System in the Czech Republic after Reform, 2009­50 139 4.8 Projected Old-Age and System Dependency Ratios in the Czech Republic, 2005­50 139 4.9 Net Replacement Rates for Male Full-Career Workers in the Czech Republic before and after Benefit Adjustment, by Income Level 141 5.1 Projected Fiscal Balance of Hungary's Public Pension System before Reform, 2000­50 150 5.2 Projected Old-Age and System Dependency Ratios in Hungary, 2005­50 150 5.3 Sources of Gross Replacement Rates in Hungary, by Income Level 166 5.4 Sources of Net Replacement Rates in Hungary, by Income Level 167 5.5 Net Replacement Rates for Male Full-Career Workers in Hungary, Europe and Central Asia, and the World, by Income Level 167 5.6 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Hungary, by Career Type and Exit Age 169 5.7 Projected Fiscal Balance of Hungary's Public Pension System after Reform, 2005­50 170 5.8 Net Replacement Rates for Male Workers in Hungary before and after Benefit Adjustment 172 x Contents 6.1 Projected Fiscal Balance of Poland's Public Pension System before Reform, 2000­50 183 6.2 Projected Old-Age and System Dependency Ratios in Poland, 2005­50 184 6.3 Sources of Gross Replacement Rates in Poland, by Income Level 200 6.4 Sources of Net Replacement Rates in Poland, by Income Level 201 6.5 Net Replacement Rates for Male Full-Career Workers in Poland, Europe and Central Asia, and the World, by Income Level 202 6.6 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Poland, by Career Type and Exit Age 203 6.7 Projected Fiscal Balance of Poland's Public Pension Scheme, 2004­50 204 7.1 Gross Replacement Rates in Romania, by Income Level 227 7.2 Sources of Net Replacement Rates in Romania, by Income Level 228 7.3 Net Replacement Rates for Male Full-Career Workers in Romania, Europe and Central Asia, and the World 228 7.4 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Romania, by Career Type and Exit Age 229 7.5 Projected Fiscal Balance of Romania's Public Pension System after Reform, 2008­50 231 7.6 Projected Old-Age and System Dependency Ratios in Romania, 2008­50 232 7.7 Net Replacement Rates for Male Workers in Romania before and after Benefit Adjustment 233 8.1 Projected Fiscal Balance of the Slovak Republic's Public Pension Scheme before Reform, 2000­50 241 8.2 Projected Old-Age and System Dependency Ratios in the Slovak Republic before Reform, 2005­50 242 8.3 Sources of Gross Replacement Rates in the Slovak Republic, by Income Level 256 8.4 Sources of Net Replacement Rates in the Slovak Republic, by Income Level 256 Contents xi 8.5 Net Replacement Rates for Male Full-Career Workers in the Slovak Republic, Europe and Central Asia, and the World 257 8.6 Net Replacement Rates for Male Middle-Income Partial-Career Workers in the Slovak Republic, by Career Type and Exit Age 258 8.7 Projected Fiscal Balance of the Slovak Republic's Public Pension Scheme after Reform, 2005­50 260 8.8 Net Male Replacement Rates in the Slovak Republic before and after Benefit Adjustment 262 9.1 Sources of Gross Replacement Rates in Slovenia, by Income Level 282 9.2 Sources of Net Replacement Rates in Slovenia, by Income Level 282 9.3 Net Replacement Rates for Male Full-Career Workers in Slovenia, Europe and Central Asia, and the World 283 9.4 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Slovenia, by Career Type and Exit Age 284 9.5 Projected Fiscal Balance of Slovenia's Public Pension Scheme, 2005­50 286 9.6 Projected Old-Age Dependency Ratio in Slovenia, 2005­50 286 9.7 Net Replacement Rates for Male Workers in Slovenia before and after Benefit Adjustment 287 Tables 1.1 Characteristics of Multipillar Pension Reforms in Transition Economies 8 1.2 Structure of Pension Systems in Eight CESE Countries 12 1.3 Taxation of Retirement Savings in Eight CESE Countries 14 1.4 Basic Pension Benefits from the Zero Pillar in Eight CESE Countries 17 1.5 Eligibility for and Benefits Provided by Old-Age Pensions in Eight CESE Countries 20 1.6 Eligibility for and Benefits Provided by Disability Pensions in Eight CESE Countries 28 1.7 Eligibility for and Benefits Provided by Survivor Pensions in Eight CESE Countries 31 xii Contents 1.8 Voluntary Pension Provisions in Eight CESE Countries 35 1.9 Health Care Provisions for Contributors and Retirees in Eight CESE Countries 36 2.1 Fiscal Balance of Bulgaria's Pension System before Reform, 1990­99 63 2.2 Structure of the Bulgarian Pension System 66 2.3 Parameters of Earnings-Related Schemes in Bulgaria before and after Reform 68 2.4 Characteristics of the Voluntary Scheme in Bulgaria 70 2.5 Eligibility Conditions for and Benefits Provided by Disability Pensions under the First Pillar Earnings-Related Scheme in Bulgaria 72 2.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in Bulgaria under the First-Pillar Earnings-Related Scheme 74 3.1 Projected Fiscal Balance of Croatia's Public Pension System before Reform, 1994­2000 90 3.2 Structure of the Croatian Pension System 93 3.3 Parameters of Earnings-Related Schemes in Croatia before and after Reform 95 3.4 Characteristics of the Voluntary Scheme in Croatia 96 3.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in Croatia 99 3.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in Croatia 101 4.1 Fiscal Balance of the Czech Republic's Pension System before Reform, 1994­2000 117 4.2 Structure of the Czech Republic's Pension System 119 4.3 Parameters of the First-Pillar Earnings-Related Scheme in the Czech Republic before and after Reform 121 4.4 Characteristics of the Voluntary Scheme in the Czech Republic 123 4.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in the Czech Republic under the First-Pillar Earnings-Related Scheme 126 4.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in the Czech Republic under the First-Pillar Earnings-Related Scheme 128 5.1 Fiscal Balance of Hungary's Pension System before Reform, 1991­96 149 5.2 Structure of Hungary's Pension System 152 Contents xiii 5.3 Parameters of Earnings-Related Schemes in Hungary before and after Reform 155 5.4 Characteristics of the Voluntary Scheme in Hungary 158 5.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in Hungary under the First-Pillar Earnings-Related Scheme 160 5.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in Hungary under the First-Pillar Earnings-Related Scheme 162 6.1 Fiscal Balance of Poland's Pension System before Reform, 1992­99 183 6.2 Structure of Poland's Pension System 186 6.3 Parameters of Earnings-Related Schemes in Poland before and after Reform 189 6.4 Characteristics of the Voluntary Scheme in Poland 192 6.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in Poland under the First-Pillar Earnings-Related Scheme 194 6.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in Poland under the First-Pillar Earnings-Related Scheme 196 7.1 Fiscal Balance of Romania's Pension System before Reform, 1992­98 213 7.2 Structure of Romania's Pension System 215 7.3 Parameters of Earnings-Related Schemes in Romania before and after Reform 217 7.4 Characteristics of Romania's Voluntary Scheme 218 7.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in Romania under the First-Pillar Earnings-Related Scheme 221 7.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions under Romania's First-Pillar Earnings-Related Scheme 222 8.1 Fiscal Balance of the Slovak Republic's Pension System before Reform, 1995­2002 241 8.2 Structure of the Slovak Pension System 244 8.3 Parameters of Earnings-Related Schemes in the Slovak Republic before and after Reform 247 8.4 Characteristics of the Voluntary Scheme in the Slovak Republic 248 xiv Contents 8.5 Eligibility Conditions for and Benefits Provided by Disability Pensions under the First-Pillar Earnings-Related Scheme in the Slovak Republic 251 8.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions under the First-Pillar Earnings- Related Scheme in the Slovak Republic 251 9.1 Fiscal Balance of Slovenia's Pension System before Reform, 1992­99 268 9.2 Structure of Slovenia's Pension System 271 9.3 Parameters of First-Pillar Earnings-Related Scheme in Slovenia before and after Reform 272 9.4 Characteristics of the Voluntary Scheme in Slovenia 274 9.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in Slovenia under the First-Pillar Earnings-Related Scheme 276 9.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in Slovenia under the First-Pillar Earnings-Related Scheme 277 Preface The former transition countries of Central, Eastern, and Southern Europe (CESE) inherited defined-benefit public pension systems financed on a pay-as-you-go basis. Under central planning, these systems exhibited fiscal strains that worsened during the early years of the transition and became unsustainable under a market economy. Recognizing that short-term fiscal pressures and incentives would worsen over the long term as a result of population aging, many CESE countries introduced reforms. Although approaches varied--particularly with regard to the choice between para- metric and systemic reforms and over the introduction of funding-- reforms typically focused on sustainability rather than benefit adequacy. At the request of--and with cofinancing from--the ERSTE Foundation, Vienna,World Bank staff prepared individual studies for eight CESE coun- tries (Bulgaria, Croatia, the Czech Republic, Hungary, Poland, Romania, the Slovak Republic, and Slovenia). The objectives were (a) to identify their motivations for reform against the backdrop of the trend toward multipillar arrangements, (b) to document their key provisions and com- pare them in the context of the World Bank's five-pillar paradigm for pen- sion reform, (c) to evaluate the sustainability and adequacy of reformed pension systems in the face of population aging, and (d) to provide a basis for recommendations to address gaps and take advantage of opportunities xv xvi Preface for further reforms. Benefit adequacy was assessed by estimating future gross and net replacement rates along both income and contribution record dimensions under steady-state conditions approximated by the year 2040. These eight studies are presented in this report. The report's introduction summarizes the case-study findings and dis- cusses several broad conclusions that emerge from them: · Fiscal sustainability has improved in most study countries, but few are fully prepared for the inevitability of population aging. · The linkage between contributions and benefits has been strength- ened, and pension system designs are now better suited to market conditions. · Levels of income replacement are generally adequate for all but some categories of workers (including those with intermittent formal-sector employment or low lifetime wages). Addressing the needs of those groups will require macroeconomic and microeconomic initiatives that go beyond pension policy. · Further reforms to cope with population aging should focus on ex- tending labor force participation by the elderly to avoid benefit cuts, which could undermine adequacy, or very high contribution rates, which could discourage formal-sector employment. · More decisive financial market reforms are needed for funded provi- sions to deliver on the return expectations of participants. These country studies were undertaken to inform policy makers, pen- sion providers, researchers, future retirees, and other stakeholders inside and outside the region about the status of future benefit adequacy in the region as well as the tasks that still lie ahead. We hope that the method- ology and comparability of analysis across countries contribute to a more informed pension reform discourse and better outcomes for the retirees of the future. Acknowledgments This report was prepared by Robert Holzmann and Ufuk Guven of the World Bank. Robert Holzmann was responsible for the overall direction of the project. He provided technical guidance and review and wrote the first chapter. Ufuk Guven wrote the eight country studies. Both authors wish to express their deep appreciation for the advice and technical input provided by Zoran Anusic, Mukesh Chawla, David Robalino, Anita Schwarz, and other World Bank staff members; the sug- gestions of Edward Whitehouse, of the Organisation for Economic Co- operation and Development (OECD) on the use of the Analysis of Pension Entitlements across Countries (APEX) model; the analytical work of Sergiy Biletsky, of the World Bank, who ran the APEX model; and the editing of Christopher Bender.They also wish to express their gratitude to the ERSTE Foundation for initiating and cofinancing the report and to its representa- tives, Rainer Munez and Karl Franz Prueller, for their seamless cooperation. The authors also thank the country experts who provided invaluable comments and suggestions on drafts of the case studies. They include Jordan Hristoskov, of the National Social Security Institute (Bulgaria); Georgi Shopov, of the Institute of Economics, Academy of Sciences (Bulgaria); Ljiljana Marusic, of the Agency for Control of Financial Services (Croatia); Jiri Kral, of the Ministry of Labor and Social Affairs xvii xviii Acknowledgments (the Czech Republic); Erzsebet Kovacs, of Corvinus University of Budapest (Hungary); Gábor Orbán, of the Economics Department of Magyar Nemzeti Bank (Hungary); Péter Holtzer, of ORIENS IM (Hungary); Agnieszka Chlon-Dominczak, of the Ministry of Labor and Social Affairs (Poland); Marek Lendack´y, of the Ministry of Finance (the Slovak Republic); and Tine Stanovnik, of the faculty of Economics and Institute for Economic Research, Ljubljana (Slovenia). Although this report was subjected to the World Bank's internal review process, the findings, interpretations, and conclusions expressed herein are those of the authors and do not necessarily reflect the views of the World Bank, its affiliated organizations, its executive directors, or the governments they represent. Abbreviations APEX Analysis of Pension Entitlements across Countries CESE Central, Eastern, and Southern Europe EET exempt-exempt-taxed EU European Union GMI Guaranteed Minimum Income GDP gross domestic product HZZO Croatian Institute for Health Insurance ILO International Labour Organization NDC notional defined-contribution NHIFA National Health Insurance Fund Administration (Hungary) NSSI National Social Security Institute (Bulgaria) PDII Pensions and Disability Insurance Institute (Slovenia) PROST Pension Reform Options Simulation Toolkit REGOS Central Registry of Insured People (Croatia) ZUS Social Insurance Fund (Poland) xix C H A P T E R 1 Introduction, Summary, and Policy Conclusions All of the former transition economies in Central, Eastern, and Southern Europe (CESE) inherited from the era of central planning traditional defined-benefit pension systems financed on a pay-as-you-go basis. Like many pay-as-you-go public pension systems elsewhere in the world, CESE pension systems were in need of reforms to address short-term fiscal imbalances and longer-term issues relating to population aging. Reforms were also needed to adjust benefit and contribution structures to meet the challenges of--as well as to take advantage of opportunities relating to--the transition to a market economy, including the wide- spread adoption of multipillar designs with improved risk-sharing across funded and unfunded pillars. By 2006, most countries in Europe and Central Asia had introduced a voluntary private pension scheme. By 2008, 14 countries--roughly half of all countries in the region--had legislated mandatory private pension schemes, and all but one of those schemes (the one in Ukraine) had been introduced. These reforms shared a number of common objectives, in particular putting the systems on a sounder finan- cial footing and better aligning them with the (very different) incentives of a market economy. Most observers would probably agree that while most countries have made progress, many of the reforms seem to have focused more on the 1 2 Adequacy of Retirement Income after Pension Reforms sustainability of the systems than on the adequacy of their retirement benefits. A pension system that delivers adequate benefits is a system that prevents old-age poverty and provides a reliable means of smoothing life- time consumption for the vast majority of the population. Indeed, this is one of the paramount objectives of pension system design. The perception that reforms placed undue emphasis on sustainability, in combination with the widespread move toward multipillar arrange- ments and funded provisions, has raised concerns about the adequacy of benefits at a time when many new retirees are receiving comparatively modest benefits. Moreover, reforms that improved but did not fully resolve issues of fiscal sustainability only heighten concerns about benefit adequacy for future generations of retirees. The population is aging rapidly in all of the CESE countries. This makes issues of sustainability and adequacy particularly important, because the inevitable consequence of the ongoing process of population aging--char- acterized by low and declining fertility rates and rising life expectancy--is that in the absence of reforms, public pension expenditures will need to rise to accommodate the larger beneficiary pool that will result from cur- rent benefit provisions and retirement ages. This is especially challenging for CESE countries with unfunded systems, because pension spending is already very high relative to gross domestic product (GDP), even though some reforms have already been enacted and the number of contributors across most age groups has fallen considerably (a drop that has not reversed, even as economic growth had picked up in some countries prior to the current financial crisis). Because many people currently of working age may not be eligible for pension benefits when they are ready to retire, gov- ernments may be compelled to consider providing them with some sort of social assistance benefits. These costs will add to the burden of paying pen- sions for those who do qualify for contributory benefits. This chapter summarizes the country-level evaluations presented in subsequent chapters of the adequacy of retirement income in eight middle-income countries in the region: Bulgaria, Croatia, the Czech Republic, Hungary, Poland, Romania, the Slovak Republic, and Slovenia. Six of these countries (Bulgaria, Croatia, Hungary, Poland, Romania, and the Slovak Republic) introduced systemic reforms involving mandatory funded pension schemes (of varying sizes and with different rules for the inclusion of current workers), together with parametric reforms to their traditional defined-benefit schemes. In Croatia, Romania, and the Slovak Republic, parametric reforms involved the introduction of a point sys- tem (described later); Poland went further, introducing a nonfinancial (or notional) defined contribution (NDC) scheme. The Czech Republic Introduction, Summary, and Policy Conclusions 3 and Slovenia are not actively working on reforms involving NDCs or mandatory funded schemes, although policy dialogue remains ongoing in both countries. This chapter is organized as follows. The first section discusses the motivation for reform across the eight countries included in the study against the backdrop of the regional (and global) trend toward multipil- lar pension arrangements. The second section summarizes the key provi- sions of the reformed systems in the eight countries within the World Bank's five-pillar framework for pension system design. The third section summarizes pension system performance against the two crucially impor- tant dimensions of adequacy and sustainability. The last section provides some policy recommendations for addressing gaps in reforms and taking advantage of further opportunities. Motivation for Reform and Policy Trends CESE countries share several common motivations for reforming their pension systems. These include the need to restore fiscal sustainability to traditional pay-as-you-go pension systems, align benefit structures, improve economic incentives, diversify risks for all parties, and (in com- mon with countries from other regions) create a vehicle for promoting financial market development (Barr and Rutkowski 2004; Holzmann 1997b; Nickel and Almenberg 2006; and Schwarz 2007). Issues of fiscal sustainability existed in many CESE countries before 1990; they were exacerbated by the transition from central planning to a market economy as a consequence of the high level of coverage under the old system (which resulted in large numbers of beneficiaries, many of whom became eligible for benefits at a relatively young age) coupled with the sharp drop in the number of contributors as a result of the initial fall in economic output, decline in labor force participation and formal employ- ment, and rise in unemployment. The level of pension expenditures in CESE countries was typically very high relative to the level of develop- ment (as measured by GDP per capita). At the same time, their capacity to collect contributions and taxes was increasingly compromised. The resulting gap between expenditures and revenues led many CESE coun- tries to consider reforms early on, but until the second half of the 1990s, fiscal pressure was accommodated largely by ad hoc measures, such as adjustments in indexation procedures and some initial parametric reforms. Prereform fiscal balances and (in some cases) projected prereform fiscal balance for the public schemes of the eight study countries indicate that deficits were invariably projected to increase over time (figure 1.1). 4 Figure 1.1 Projected Pension System Fiscal Balances before Reform in Eight CESE Countries Bulgaria Croatia Czech Republic Hungary 20 20 20 20 15 15 15 expenditure 15 expenditure expenditure GDP 10 GDP 10 revenue GDP 10 GDP 10 of of of revenue of 5 revenue 5 5 5 entagec 0 entagec 0 entagec 0 entagec 0 per per per per ­5 2001 2005 2010 2015 2020 2030 2035 2040 2045 2050 ­5 1994 1995 1996 1997 1998 1999 2000 ­5 1994 1995 1996 1997 1998 1999 2000 ­5 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 deficit deficit deficit deficit ­10 ­10 ­10 ­10 Poland Romania Slovak Republic Slovenia 20 20 20 20 expenditure expenditure 15 15 15 15 expenditure GDP 10 GDP 10 GDP revenue GDP 10 10 of revenue of of revenue of revenue 5 5 expenditure 5 5 entagec 0 entagec 0 entagec 0 entagec 0 per per per per ­5 2000 2010 2020 2030 2040 2050 ­5 1992 1993 1994 1995 1996 1997 1998 ­5 2000 2010 2020 2030 2040 2050 ­5 1992 1993 1994 1995 1996 1997 1998 1999 deficit deficit deficit deficit ­10 ­10 ­10 ­10 Sources: European Commission 2007, national ministries, and national social security institutions. For details, see individual country chapters. Note: Projections of prereform expenditures and revenues were not available for Hungary; historical data are provided. Introduction, Summary, and Policy Conclusions 5 The long-term deterioration expected in the fiscal balances of the pen- sion systems of the eight study countries will ultimately be driven by fur- ther population aging, a phenomenon common to all countries in the region. Projections of old-age dependency ratios (that is, the ratio of the population age 65 and older to the population age 15­64, a good proxy for the impact of aging on pay-as-you-go pension schemes) for six groups of countries is shown to highlight the relative magnitude of aging in the region (figure 1.2). These projections show that while aging in Central Europe and the Baltic region and in South Eastern Europe is currently less pronounced than it is in the EU15 countries (members of European Union before 2004), the rate of aging is higher, so that by 2050 the old- age dependency ratio in Central Europe and the Baltic region is expected to surpass that of the EU15, more than doubling in less than 50 years. For many of the countries in the region, old-age dependency ratios actually underestimate the impact of aging on their pay-as-you-go pension sys- tems, because retirement ages in most of these countries are well below 65 and the number of pension system contributors, which declined dur- ing the transition, has shown no indication of returning to anywhere close to pretransition levels. Figure 1.2 Old-Age Dependency Ratios, 2000­50, by World Region 0.6 0.5 tioar y 0.4 0.3 dependenc 0.2 old-age 0.1 0.0 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 CEB EU15 SEE North world South Source: Authors' estimates based on UN 2007. Note: The old-age dependency ratio is the ratio of the population age 65 and older to the population age 15­64. CEB = Central Europe and the Baltics; EU15 = members of European Union before 2004; SEE = Southeastern Europe. 6 Adequacy of Retirement Income after Pension Reforms In addition to the need to address issues of short- and long-term fiscal pressure, another common motivation for reform in CESE countries was to better allow their pension systems to function in a market economy. Their inherited pension systems shared a number of common fea- tures, including the use of unfunded (pay-as-you-go) financing based on contributions levied on wages; benefit formulas based on wages at retirement, with little linkage to lifetime contributions and often with a redistributive objective intended to support low-income earners; low retirement ages; and many privileges for special groups, despite the fact that most CESE countries had a single scheme that also covered civil ser- vants and farmers. The special treatment given to many groups and the structure of benefits may have been conceptually aligned with public own- ership of enterprises and centralized contribution payments. It became increasingly dysfunctional in a market economy with the privatization of large state enterprises and the emergence of small and medium-size enter- prises and the self-employed. Moreover, the use of pay-as-you-go financ- ing placed all risk on plan sponsors--that is, governments--which were also faced with the rapid aging of their populations. For their part, individ- uals were deprived of the opportunity to profit from the diversification of risk and the investment of their savings in emerging financial sectors. At the beginning of economic transition, the financial sectors in CESE countries consisted only of state-owned banks. These banks catered to public enterprises and were essentially an arm of the central planning process. The financial instruments available to individuals and small enter- prises were limited primarily to cash, often held in foreign currencies, and savings accounts yielding low nominal returns. Although the reform of banking systems (including bank privatization) and the establishment of insurance and securities markets were part and parcel of the reform process in all CESE countries, the development of financial systems takes time. Even today, the financial sectors of many CESE countries are less developed than those in countries elsewhere with similar income. This recognition contributed to the consideration of reforms, including the introduction of funded pension pillars, that were expected to accelerate financial market development, as they did in Chile (Holzmann 1997a). Against this backdrop, all countries in the region initiated a process of pension reform motivated by the need to reform their existing systems and, in many (but not all) cases, by the trend toward multipillar structures, which started in Latin America. The publication of the World Bank's sem- inal report Averting the Old-Age Crisis (World Bank 1994), motivated in part by the reform challenges faced in Latin America, supported this Introduction, Summary, and Policy Conclusions 7 trend. After reviewing the limited alternatives then being proposed by the literature--and then and now by the International Labour Organization-- many reformers concluded that a more radical approach, including a move toward multipillar systems with mandatory fully funded, defined- contribution pension schemes, was required.1 Several transition economies have introduced multipillar pension sys- tems. Hungary and Kazakhstan were the first to do so, in 1998. By 2008, 13 countries in the region had introduced funded pillars, with Ukraine condi- tionally scheduled to follow in 2010 or 2011. All CESE countries have undertaken parametric reforms, some significant, others basic. Some coun- tries, including the Czech Republic and Slovenia, have resisted introducing mandated funded pillars; their pay-as-you-go schemes require additional parametric reforms to become sustainable. In Armenia, Montenegro, and Serbia, the debate over funded pillars continues.Albania,Azerbaijan, Bosnia and Herzegovina, the Kyrgyz Republic, and Turkmenistan have yet to undertake major reforms; they may need to defer consideration of funded pillars until preconditions have been met. The countries that have undertaken multipillar reforms may have been inspired by the examples of Chile and other Latin American countries, but each of them has taken its own approach (table 1.1). Of the 14 coun- tries that have legislated reforms, 12 have retained a main pay-as-you-go (first) pillar scheme. Mandatory funded (second-) pillar schemes supple- menting first-pillar schemes are expected to diversify risk while providing roughly half of retirement income. The decision to retain first-pillar schemes was driven primarily by the major financing needs a full transi- tion, such as Chile and Mexico implemented, would have called for. The institutional arrangements for private pension funds vary across CESE countries; in most cases, they diverge from the Latin American examples with regard to sponsoring institutions and supervision. A num- ber of CESE countries have taken innovative approaches to reforming their first-pillar schemes and have tried to learn from the experiences of Latin American countries in keeping the costs and fees of their funded second pillars low. The first-pillar reforms fully introduced in Latvia and Poland and partially introduced in the Russian Federation were inspired by the example of Sweden, which pioneered the use of NDCs, which mimic defined contribution schemes while remaining largely unfunded (see Holzmann and Palmer 2006). The introduction of a point system in Croatia, Romania, Serbia, the Slovak Republic, and Ukraine was inspired by the German and French systems. This system behaves like an NDC scheme but lacks many of its strengths. 8 Table 1.1 Characteristics of Multipillar Pension Reforms in Transition Economies Size of second pillar Projected pension Share of workforce in (percentage fund assets in 2020 funded pillar in 2008 Switching of rules Country Starting date First (or zero) pillar of payroll) (percentage of GDP) or earlier (percent) to new system Bulgaria January 2002 Pay-as-you-go defined 5.00 ­25 70 Mandatory for benefit workers under age 42 Croatia January 2002 Pay-as-you-go defined 5.00 25 80 Mandatory for workers under benefit age 40, voluntary for workers age 40­50 Estonia July 2002 Pay-as-you-go defined 6.00 20 75 Voluntary benefit (optout + 2) Hungary January 1998 Pay-as-you-go defined 8.00 32 45 Mandatory for new entrants, benefit voluntary for others Kazakhstan January 1998 Basic pension 10.00 35 82 Mandatory Kosovo January 2002 Universal/minimum 10.00 8 30 Mandatory consumption basket level Latvia July 2001 (NDC Pay-as-you-go defined 4.00, growing 25­30 72 Mandatory for workers under January 1996) contribution/ to 10.00 age 30, voluntary for workers nonfinancial defined by 2010 age 30­50 contribution Lithuania January 2004 Pay-as-you-go defined 5.50 35­40 55 Voluntary benefit Macedonia January 2006 Pay-as-you-go defined 7.12 26 25 Mandatory for new entrants benefit Poland January 1999 Pay-as-you-go defined 7.3 34 70 Mandatory for workers under 30, contribution/ voluntary for workers nonfinancial defined 30­50 contribution Romania Registration Pay-as-you-go defined 2 in 2008, 9 65 Mandatory for workers under 35, completed; benefit growing voluntary for workers contributions gradually 36­45 began June 2008 to 6 by 2016 Russia January 2002 Pay-as-you-go defined (6 in 2008) -- 33 Mandatory for contribution/ workers under 50 nonfinancial defined contribution Slovak January 2005 Pay-as-you-go defined 9 20 75 Mandatory for new entrants Republica benefit Ukraine July 2009 or Pay-as-you-go defined 2, growing 16 -- Mandatory for new entrants January 2010 benefit to 7 Source: Holzmann 2009. -- Not available. Note: Data are as of January 2008. Systems are operating in all countries except Ukraine, where reforms have been partially legislated. a. Made optional for new entrants in January 2008; participants were given six-month window to opt out (or in) of the second pillar. 9 10 Adequacy of Retirement Income after Pension Reforms The only CESE countries to follow Chile's approach to pension reform are Kazakhstan and Kosovo. Both countries rely exclusively on a basic (zero) pillar (a noncontributory scheme intended to provide a minimal level of income protection) and a mandated funded (second) pillar. In Kazakhstan, all workers were enrolled in the new scheme, although their rights under the old pay-as-you-go scheme were recognized (Hinz, Zviniene, and Vilamovska 2005). In Kosovo, after the conflict the new authorities had neither the means nor the records to recognize accrued rights. A special feature of the Kosovo scheme is that all assets are invested internationally, because the domestic market is not yet consid- ered ready for local investing (Gubbels, Snelbecker, and Zezulin 2007). Characteristics of Reformed Pension Systems An accurate assessment of the adequacy and fiscal sustainability of a pen- sion system must start from a clear understanding of its design. This sec- tion summarizes the key provisions of the reformed pension systems in the eight study countries as of January 2008. These provisions include the structure of the individual pillars of social insurance; the rules governing pension taxation; institutional structure; coverage; and the rules govern- ing old-age, disability, and survivorship pensions. These provisions are then discussed within the framework developed by the World Bank, which generally recommends including a funded component if conditions are appropriate but which increasingly recognizes that a range of choices is available to policy makers to provide effective old-age protection in a manner that is fiscally responsible. The World Bank suggests that pension systems be composed of some combination of five basic pillars: · a noncontributory (or zero) pillar (in the form of a demogrant, social pension, or social assistance benefit) intended to provide a minimal level of income protection; · a first-pillar contributory system linked to earnings, which seeks to re- place a portion of preretirement income; · a mandatory second pillar (essentially, individual savings accounts), which can be designed in various ways; · a voluntary third pillar, which is flexible and discretionary (this pillar, too, can take a variety of forms); and · a fourth pillar of informal intrafamily or intergenerational sources of financial and nonfinancial support to the elderly, including access to health care and housing (Holzmann and Hinz 2005). Introduction, Summary, and Policy Conclusions 11 Pillar Design All of the reformed pension systems of the eight study countries provide old-age income support through some combination of all five of these pillars (table 1.2). All countries have a zero pillar, the purpose of which is to alleviate poverty among the elderly. In most countries, the zero pil- lar is part of a broader scheme of social assistance available to everyone, regardless of age, intended to guarantee a minimum income level. Three countries (Bulgaria, Hungary, and Slovenia) provide an age-related social pension specifically for the elderly. Zero pillars are all means tested (to target lower-income groups), universally publicly managed, noncontribu- tory, and financed with general tax revenue. The amount of the benefit is typically adjusted to ensure that total household income meets some minimum state-defined level, which is often related to other forms of assistance and adjusted for inflation (or wage growth) on an ad hoc basis. The relation between the income level and the poverty line is often ten- uous, and poverty lines vary widely across countries. All eight countries reformed their existing pay-as-you-go, first-pillar schemes, all of which are earnings related in the sense that benefits in retirement depend, in varying degrees, on earnings received and contribu- tions paid while working. The structure of their reformed schemes differs across countries. Bulgaria, the Czech Republic, Hungary, and Slovenia rely on a traditional defined-benefit design in which benefits depend on (a) some measure of assessed income, (b) an annual accrual rule (the per- centage of assessed income that is replaced in benefits for each year of contributory service), and (c) the length of contributory service. In con- trast, Croatia, Romania, and the Slovak Republic replaced their defined- benefit schemes with a system based on points. When appropriately implemented, such a system is functionally equivalent to a reformed defined-benefit system in which lifetime income is revalued relative to the average wage. Poland replaced its defined-benefit scheme with an NDC scheme that functionally mimics a funded defined-contribution sys- tem while remaining financed on a pay-as-you-go basis (plus a reserve fund). Contributions to the scheme are earmarked to individual accounts and remunerated with (notional) interest. At retirement, account bal- ances in combination with conditional life expectancy are used to deter- mine initial benefits. Provisions for the remaining pillars are as follows: · Six of the eight countries (Bulgaria, Croatia, Hungary, Poland, Romania, and the Slovak Republic) reduced first-pillar benefits for future 12 Table 1.2 Structure of Pension Systems in Eight CESE Countries Zero pillar First pillar (mandated, Fourth pillar (noncontributory) earning related) Second pillar (mandated, Third pillar (voluntary) (health care) Provision Type Function Provision Function earnings related) Provision Function Provision Function Country (Public) (MBB) (Poverty) (Public) Type (Insurance) Provision Type Function (Private) Type (FDC) (Insurance) (Public) (Insurance) Bulgaria X X X X NDB X Private FDC Insurance X X X X X Croatia X X X X NPS X Private FDC Insurance X X X X X Czech X X X X NDB X n.a. n.a. n.a. X X X X X Republic Hungary X X X X NDB X Private FDC Insurance X X X X X Poland X X X X NDC X Private FDC Insurance X X X X X Romania X X X X NPS X Private FDC Insurance X X X X X Slovak X X X X NPS X Private FDC Insurance X X X X X Republic Slovenia X X X X NDB X n.a. n.a. n.a. X X X X X Source: Unpublished World Bank pension database. Note: MBB: means-tested basic benefit; NDB: nonfinancial defined benefit; NPS: nonfinancial point system; NDC: nonfinancial (notional) defined contribution; FDC: financial defined contribution. n.a. Not applicable. X Countries have these pillars. Introduction, Summary, and Policy Conclusions 13 beneficiaries and complemented their first-pillar schemes with man- dated earnings-related, funded second-pillar schemes. These schemes are defined-contribution schemes that rely on privately managed pen- sion funds for administration and asset management. · All eight countries introduced voluntary funded, third-pillar schemes to provide individuals with a mechanism for supplementing the benefits paid by the mandatory pillars. These schemes rely on private-sector financial institutions--such as insurance companies, mutual funds, and pension funds--for administration and asset management. · All eight countries provide health insurance on a contributory basis to the active population. Health insurance schemes extend to retirees receiving public pensions. Access to this fourth pillar is crucially important for the design of a pension system and its target levels of income replacement. Taxation of Contributions and Benefits The taxation of contributions and benefits for each pillar of a pension sys- tem has major bearing on the adequacy of pension benefits and the degree to which take-home pay is actually replaced in retirement. The difference between gross replacement rates (the ratio of benefits disbursed to pretax preretirement earnings) and net replacement rates (the ratio of benefits actually received to posttax preretirement earnings) is usually substantial. What matters for the elderly is the amount of their net pension, because it is on this basis that they finance their consumption in retirement. Taxation of pension contributions, investment returns (during the accu- mulation phase of funded schemes), and benefits upon disbursement range widely across the eight study countries and across the individual pil- lars of their pension systems (table 1.3). Among pension policy experts, there is consensus that earnings- related schemes should be subject to some form of taxation; there is less agreement on whether earnings-related schemes should be subject to comprehensive income taxation or expenditure (consumption) taxation. Such taxation is generally considered to be less distortionary with regard to savings decisions (decisions relating to whether to consume now or in the future). Certain principles apply to the taxation of retirement "savings" (box 1.1). These principles are typically applied to funded schemes but should also be applied to unfunded schemes. For noncontributory zero-pillar schemes, the first two opportuni- ties for taxation do not exist, and means-tested benefits in all study 14 Table 1.3 Taxation of Retirement Savings in Eight CESE Countries First pillar (earnings related) Second pillar (earnings related) Third pillar (voluntary) Fourth pillar (health care) Investments/ Investments/ Investments/ Investments/ Country Contributions capital gains Benefits Contributions capital gains Benefits Contributions capital gains Benefits Contributions Capital Gains Benefits Bulgaria Exempt Exempt Exempt Exempt Exempt Exempt Exempta Exempt Taxed Exempt n.a. Exempt Croatia Exempt Exempt Taxed Exempt Exempt Taxed Exemptb Exempt Taxed Exempt n.a. Exempt Czech Republic Taxedc Exempt Exemptd n.a. n.a. n.a. Exempte Exempt Taxed Taxed n.a. Exempt Hungary Taxed Exempt Exempt Taxed Exempt Exemptf Exemptg Exempt Exempth Exempt n.a. Exempt Poland Exempt Exempt Taxed Exempt Exempt Taxed Taxedi Exemptj Exempt Exempt n.a. Exempt Romania Exempt Exempt Taxed Exempt Exempt Taxed Exemptk Exempt Taxed Exempt n.a. Exempt Slovak Republic Exempt Exempt Exemp Exempt Taxed Exempt Exemptl Taxed Taxed Exempt n.a. Exempt Slovenia Exempt Exempt Taxed n.a. n.a. n.a. Exemptm Exempt Taxed Exempt n.a. Exempt Source: European Commission 2007, national ministries, and national social security institutions. For details, see the individual country chapters. Note: All benefits under the noncontributory (zero) pillar are exempt from taxation. n.a. Not applicable. a. Contributions up to 10 percent of earnings are exempt from the taxation. Contributions up to lev 60 per worker per month paid by employers are exempt from the corporate income tax base. b. Contributions up to HRK 12,000 a year can be deducted from personal income for tax purposes. c. Since January 1, 2008, contributions paid to the first pillar are part of the income tax base, consistent with the tax reform that introduced a flat tax of 15 percent. d. Pensioners are provided with a large tax allowance on pension income. Pensioners with total taxable income of less than 80 percent of average earnings do not pay income taxes, exempting pensions from taxation for all pensioners except those with substantial supplementary income. Pensioners who are taxed pay a rate lower than that levied on earnings. Pensions thus pay sub- stantially less in taxes than workers with the same total income. e. Contributions of CZK 6,000­2,000 and employer contributions up to 5 percent of wage are tax exempt. f. There is great uncertainty regarding the taxation of benefits after 2013. g. Thirty percent of contributions are tax deductible up to an annual cap of HK$100,000. h. Benefits are exempt from tax if taken as a qualified annuity and the accumulation period is at least 20 years. If the accumulation period is 10­20 years, benefits are partially taxed. i. Employer contributions to the third pillar are deductable from the employer's taxable income. j. Employees are granted tax relief up to 150 percent of the average wage, above which they must pay taxes for capital gains and retirement savings. k. Up to EURO 200 per year per participant is tax exempt. l. Up to Sk 12,000 annually is exempt from taxation. m. Up to 24 percent of contributions to the first pillar are tax exempt. Introduction, Summary, and Policy Conclusions 15 Box 1.1 Taxation of Retirement Savings Like other forms of savings, retirement savings can be taxed when contributions are made; as investment income and capital gains accrue (except, of course, in pay-as-you-go pension schemes); or when benefits are actually paid. Four of the eight possible combinations are examined in the table below for a hypothetical contribution of 100 units of currency made five years before a worker's retire- ment. Investments are assumed to earn a 10 percent annual rate of return; the tax rate is assumed to be 25 percent. Table Retirement Benefits Paid under Various Tax Regimes Exempt- Taxed- Taxed- Exempt- exempt-taxed taxed-exempt exempt- taxed-taxed Item (EET) (TTE) exempt (TEE) (ETT) Initial contribution 100 100 100 100 Taxes levied on contribution 0 2 25 0 Initial account balance 100 75 75 100 Investment returns net of taxes 61 46 33 44 Ending account balance 161 121 108 144 Taxes paid on distributions 40 0 0 36 Pension net of taxes 121 121 108 108 Source: Whitehouse 1999. In this stylized illustration, the first two regimes--exempt-exempt-taxed (EET), in which contributions and investment income are exempt from taxation but benefits are taxed, and taxed-exempt-exempt (TEE), in which contributions are taxed but investment income and benefits are exempt from taxation--provide the same level of postretirement income. Both also provide the same present value of tax revenues, although the revenues under EET (a classic expenditure tax) are deferred until the worker retires whereas revenues under TEE (a prepaid expendi- ture tax) are received earlier. The remaining two regimes (both comprehensive income taxation) provide equivalent levels of postretirement income and tax rev- enues. Relative to expenditure taxes, however, they yield more taxes and lower overall rewards for saving. 16 Adequacy of Retirement Income after Pension Reforms countries are low enough that they are not subject to income taxes, making zero-pillar schemes effectively fully exempt from taxation in the eight study countries.2 Taxation of first-pillar schemes varies. Bulgaria and the Slovak Republic fully exempt benefits from taxation; Croatia, Poland, Romania, and Slovenia have classic expenditure taxes. The Czech Republic and Hungary have prepaid expenditure taxes. All countries but one that have introduced mandatory second pillars apply the same rules to their sec- ond-pillar schemes as they apply to their first-pillar schemes. The excep- tion is the Slovak Republic, where the first pillar is fully exempt from taxation but investment income and capital gains under the second pillar are taxed. Taxation of voluntary third-pillar schemes varies widely across the eight study countries (as well as with respect to the way in which first- and second-pillar schemes are taxed). All eight countries exempt contribu- tions, investment income, and capital gains (except the Slovak Republic); all countries except Hungary and Poland tax benefits (consistent with an expenditure tax approach). Because voluntary schemes are more popular with higher-income groups, the exemption of benefits is regressive. All countries except the Czech Republic tax health care contributions; all countries exempt health care benefits from taxation. Benefits from Zero-Pillar Schemes All eight countries provide minimum old-age benefits, aimed at alleviating poverty among the elderly. In most countries benefits are part of a broader noncontributory program of social assistance available to everyone, regard- less of age, intended to guarantee a minimum income level, independent of other individual characteristics (table 1.4). Such means-tested programs were necessary in transition economies, because only certain categories (such as people with disabilities) were eligible for social assistance under central planning; everyone else had access, in principle, to paid employ- ment followed by a pension. Once central planning was abandoned, this was no longer the case.3 With the abolition of most categorical benefits and the introduction of a guaranteed minimum income level for everyone, most CESE countries have been reluctant to address the problem of poverty among the elderly with social pensions. Social pensions provide an income guarantee, but they apply only to people of a certain age. Moreover, while benefits are typically higher than general social assistance benefits, means-testing is typically less intrusive. Table 1.4 Basic Pension Benefits from the Zero Pillar in Eight CESE Countries (continued) Total expendi- ture (percent- Country Benefit Coverage/Eligibility Year Benefit level Indexation Number of beneficiaries age of GDP) Bulgariaa Social pen- Individuals above age 70 who 2006 63 leva (17.75 50 percent 4,592 (0.2 percent of total -- sion are not collecting a pension; percent of average prices, 50 pensioners) in 2005 average income per family wage) percent member must be lower than wage the guaranteed minimum growth income for full 12-month previous period year Croatia Guaranteed Individuals with income below n.a. Percentage of state- Ad hoc 2.7 percent of population in 0.22 minimum guaranteed minimum income defined subsis- 2005 income tence allowance Czech Guaranteed Individuals with income below 2006 3,126 koruny Prices About 4 percent of households -- Republic minimum guaranteed minimum income income Hungary Old-age Individuals age 62 and older n.a. Supplements actual Old-age 6,679 beneficiaries (0.4 percent 0.01 allowance with income below 80 percent income to reach 80 minimum of population age 62 of minimum old-age pension percent of old-age pension and older) in 2003 minimum pension Polandb Guaranteed Individuals with income below 2006 About 16 percent of Regular -- -- minimum guaranteed minimum income average wage increases income based on social assistance 17 legislation (continued) 18 Table 1.4 Basic Pension Benefits from the Zero Pillar in Eight CESE Countries (continued) Total expendi- ture (percent- Country Benefit Coverage/Eligibility Year Benefit level Indexation Number of beneficiaries age of GDP) Romania Guaranteed Individuals with income below 2006 RON 92 (9 percent Changes in 834,000 beneficiaries in 2005 0.2 minimum guaranteed minimum income of average wage) consumer income price index Slovak Guaranteed Individuals with income below 2006 Sk 4,980 (about 27 Minimum 182,479 beneficiaries in 2007c 0.45 Republic minimum guaranteed minimum income percent of average subsistence income wage) level (close to con- sumer price index) Slovenia State Individuals 65 and older who n.a. One-third of mini- Growth of -- -- pension do not qualify for a pension mum pension minimum from the first-pillar pension assessment base pension schemed assessment base Source: European Commission 2007, national ministries, and national social security institutions. For details, see individual country chapters. -- Not available. a. Another noncontributory source of income support for the elderly comes from the guaranteed minimum income, which provides a means-tested benefit. b. There is also a minimum pension guarantee under the old-age pension system. For pensioners who contributed for at least 25 years (men) or 20 years (women) whose total pension falls below a certain threshold, the difference is topped up from the state budget. In 2008 the guaranteed benefit was Zl 636 per month. Minimum pensions are taxed according to the general PIT (Personal Income Tax) taxation rules. c. About 21.2 percent of all beneficiaries (38,606 people) are of pensionable age. d. Beneficiaries must also have lived in Slovenia for at least 30 years between the ages of 15 and 65. Introduction, Summary, and Policy Conclusions 19 Three countries (Bulgaria, Hungary, and Slovenia) offer social pensions in addition to the general social assistance scheme. In Bulgari,a social pen- sions are provided to individuals age 70 and older who are not collecting old-age or disability pensions. Eligibility is based on average income per family member. The allowance is means tested and adjusted in value such that the beneficiary's total income reaches the minimum threshold (roughly 18 percent of the average wage). In Hungary, eligibility is lim- ited to individuals age 62 and older who can demonstrate that their total income falls below 80 percent (95 percent for couples) of the minimum old-age pension. In Slovenia, eligibility is limited to individuals age 65 and older who have lived in Slovenia for at least 30 years and who do not qualify for an old-age pension. Benefits are equal to one-third of the min- imum pension assessment base. Benefits from Earnings-Related First- and Second-Pillar Schemes Earnings-related, first-pillar pension schemes in all eight study countries provide old-age pensions, disability pensions, and survivorship benefits. Second-pillar schemes are structured primarily to provide old-age pen- sions only. The key characteristics of each of these types of benefits across all eight countries are discussed in this section. Old-age benefits. Earnings-related, first-pillar pension schemes--and (for countries that have them) second-pillar schemes--exhibit both similarities and differences in terms of eligibility conditions and benefit provisions across the eight study countries (table 1.5). The most signif- icant of these conditions and provisions are vesting periods, contribution rates, contribution ceilings, benefit calculations, retirement ages, and benefit indexation. VESTING PERIODS. Vesting periods (that is, the minimum period of con- tributory service needed to qualify for a pension upon reaching the min- imum or normal retirement age) are typical of traditional defined-benefit schemes. They serve to prevent a form of arbitrage in which people work for only a few years to become eligible for a minimum (in some cases, a flat) benefit worth far more than the contributions they paid toward their benefits. In actuarially fair pension systems (that is, systems in which the lifetime value of a person's benefits is, by design, roughly equal on aver- age to the lifetime value of his or her contributions), vesting periods are unnecessary, except for determining eligibility for a minimum pension. To 20 Table 1.5 Eligibility for and Benefits Provided by Old-Age Pensions in Eight CESE Countries Contribution Contribution Pension assessment Benefit Country Pillar Vesting period rate ceiling Benefit rate base Retirement age indexation Bulgariaa First 15 years TCR: 23.00 Fixed annually 1 percent per year Highest 3 of last 63 for men; gradually 50 percent percent (14.95 in budget 15 years before 1997 increasing to 60 for prices, percent (1,400 leva in plus entire working women by 2009 50 percent by employer, 2007, roughly period afterward wage growth 8.05 percent 350 percent of of previous by employee) average wage) year Second No minimum PCR: 5 percent Fixed annually Varies depending on Accumulated funds 63 for men; gradually Income stream (for people in budget life expectancy at increasing to 60 for from conver- born after (1,400 leva in retirement and rate women by 2009 sion of capital 1959) 2007) of return accumulation Croatia First 15 years TCR: 20 percent 5 times average 0.75 percent in 2008, Flat rate plus Swiss- 60 for women, 65 for 50 percent (all from wage decreasing over time indexed lifetime men prices, employee) (0.25 percent of earnings for second- 50 percent average wage and pillar participants wages 25.00 percent of Swiss- indexed point value for second-pillar participants) Second 15 years PCR: 5 percent 5 times average Varies depending on life Accumulated funds 60 for women, 65 for Price-indexed (all from wage expectancy at retire- men annuity employee) ment and rate of return Czech First 35 years at normal TCR: 28.00 None Flat benefit plus Gradually increasing to By 2030 65; for all men Minimum of Republic retirement age percent (21.5 1.5 percent per year average of last and for women with prices plus by 2019; 20 years percent by 30 years'earnings no or one child, one-third of for people who employer, 6.5 by 2016 64 for women with 2 real wage retire 5 or more by employee) children, 63 for growth years after women with 3 normal children, and 62 for retirement age women with more than 3 children Hungary First 15 years at age 62 TCR: 33.5 per- Employee: Set Until 2013: 33.00 Average lifetime 62 for men; 62 for 50 percent under special cent (24.0 annually by percent for first 10 earnings revalued by women by 2009 prices, (strict) condi- percent government, years, 2.00 percent for wage growth 50 percent tions; 20 years by employer, at about 11­25.00 years, 1.00 wages normally 9.5 percent 8 times percent for 26­36 by employee) minimum years, and 1.50 wage percent beyond Employer: No 36 years maximum After 2013: 1.65 percent per year Second No minimum PCR: 8 percent Employee: Set Varies depending on life Accumulated funds 62 for men; 62 for 50 percent (all from annually by expectancy at women by 2009 prices, employee) government, retirement and rate 50 percent at about eight of return wage, times mini- indexed mum wage annuity Employer: No maximum 21 (continued) 22 Table 1.5 Eligibility for and Benefits Provided by Old-Age Pensions in Eight CESE Countries (continued) Contribution Contribution Pension assessment Benefit Country Pillar Vesting period rate ceiling Benefit rate base Retirement age indexation Polandb First No minimum; PCR: 19.52 per- 2.5 times Varies depending on life Notional capital accu- Gradually increasing Mixed price- individuals are cent (9.76 per- national expectancy at retire- mulation to 65 for men by wage eligible for mini- cent by average wage ment and notional rate 2014 and 60 for formula, in mum pension employer, 9.76 of return women by 2009 which wages after 20 years percent by account for (women) and employee) 20 percent of 25 years (men) indexation Second None; individuals PCR: 7.3 percent 2.5 times Varies depending on life Capital accumulation Gradually increasing Price-indexed are eligible for (all from national expectancy at to 65 for men by annuity minimum pen- employee) average wage retirement and rate 2014 and 60 for sion after of return women by 2009 20 years (women) and 25 years (men) Romaniac First 15 years TCR: 29.75 per- None Based on number of Lifetime average Gradually increasing to Adjusted based cent (20.25 wage-indexed points indexed to nominal 65 for men and 60 for on changes in percent by earned wage growth women by 2015 point value employer, (cannot fall 9.50 percent below 45 per- by employee) cent of aver- age wage) Second Not established by PCR: 2 percent None Varies depending on life Capital accumulation Gradually increasing to Specific regula- new law increasing to expectancy at retire- 65 for men and 60 for tion does not 6 percent over ment and rate of return women by 2015 yet exist period of 8 years Slovak First pillar 15 years PCR: 28.75 per- 3 times average 1.19 percent per year Lifetime average 62 for men; gradually 50 percent Republic cent:18.00 per- wage (based on number of indexed to nominal increasing to 62 for prices, 50 per- cent old-age wage-indexed points wage growth women by 2016 cent nominal (4.00 percent earned) wage growth by employee, 14.00 percent by employer); 6.00 percent disability (3.00 percent by employee, 3.00 percent by employer); 4.75 percent reserve fund (100.00 percent by employer) Second 10 years PCR: 9 percent 3 times average Varies depending on life Accumulated funds 62 for men; gradually Depends on (all old-age) wage expectancy at increasing to 62 for options (all from retirement and rate women by 2015 chosen employee) of return 23 (continued) 24 Table 1.5 Eligibility for and Benefits Provided by Old-Age Pensions in Eight CESE Countries (continued) Contribution Contribution Pension assessment Benefit Country Pillar Vesting period rate ceiling Benefit rate base Retirement age indexation Slovenia First 15 years TCR: 24.35 No maximum 35 percent for men, Gradually increased to Gradually increasing to Wage growth percent (15.5 38 percent for women best 18 years in 2008 63 for men by 2009 percent for first 15 years of con- and 61 for women by employee, 8.85 tribution; 1.5 percent 2023 percent per year beyond employer) 15 years Sources: European Commission 2007, national ministries, and national social security institutions. For details, see the individual country chapters. Note: TCR = total contribution rate (including contributions to both the first and the second pillar, where applicable); PCR = pillar-specific contribution rate; n.a. = Not applicable. a. In addition to the main (universal) defined = contribution scheme, there are also fully funded defined = contribution schemes as part of the second pillar. b. The total contribution rate is 27.92 percent (19.52 percent for old-age pensions, 6.00 percent for disability pensions, and 2.45 percent for sickness and maternity benefits). Individuals participating in only the first pillar pay 19.52 percent to the first pillar for old-age pensions (split equally between employers and employees). Individuals participating in both the first and the second pillars pay 12.22 percent to the first pillar (9.76 percent paid by employers and 2.46 percent paid by employees) and 7.3 percent to the second pillar (paid entirely by employees). Employers pay contributions for work injury. The rate varies by industry. c. Total contribution rate will be reduced to 28 percent (18.5 percent by employer, 9.5 percent by employee) in 2009. Introduction, Summary, and Policy Conclusions 25 avoid the payment of pensions on small amounts, lump-sum payments can be provided. Of the eight study countries, only Poland--with its combination of an NDC first-pillar scheme and a defined contribution second pillar--has an actuarially fair system. As a result, it is the only country not to specify a vesting period, although eligibility for its minimum pension is subject to a minimum contributory period of 20 years for women and 25 years for men. In all other study countries, vesting periods for first-pillar benefits are 10, 15, 20, or 25 years. Some countries (such as Croatia and the Slovak Republic) but not oth- ers impose vesting periods on second-pillar benefits. Because defined con- tribution schemes are, by construction, actuarially fair, vesting periods for second-pillar schemes are theoretically unnecessary; it is possible that they are imposed for administrative reasons or out of the desire for con- sistency with first-pillar rules. In practice, vesting periods tend to reduce incentives to enroll in voluntary schemes as well as the effectiveness of an actuarially fair structure with regard to labor supply decisions. CONTRIBUTION RATES. Contribution rates for first-pillar schemes in most of the eight study countries collectively cover the cost of old-age and dis- ability pensions as well as survivorship benefits. The exceptions are Poland, where old-age pensions have an earmarked contribution rate of 19.52 percent, and the Slovak Republic, where the contribution rate is 18 percent. Contribution rates in the other seven countries range from 20.0 percent (in Croatia) to 28.0 percent (in the Czech Republic) and 24.35 percent (in Slovenia), neither of which has a second-pillar scheme. Contribution rates for second-pillar schemes range from 5 percent (in Bulgaria and Croatia) to 9 percent (in the Slovak Republic). In most countries they are being phased in gradually. In Romania, for example, the contribution rate, initially set at 2 percent, is gradually being increased such that it will reach 6 percent by 2016. Assuming that the contribu- tion required to fund disability benefits is roughly 6­8 percentage points of the total pension levy, second-pillar contribution rates in the study countries are roughly a third the size of first-pillar contribution rates for old-age pensions (including survivorship benefits). CONTRIBUTION CEILINGS. Contribution ceilings differ widely across the study countries. Three countries (the Czech Republic, Romania, and Slovenia) have no ceilings. In three countries the ceiling is set as a multiple of the average wage (5 times the average wage in Croatia, 3 times the 26 Adequacy of Retirement Income after Pension Reforms average wage in the Slovak Republic, and 2.5 the average wage in Poland). In Bulgaria and Hungary, the government sets the ceiling annually. Contribution ceilings limit the scope of a public pension scheme's mandate and create space for higher-income workers to diversify their retirement savings outside of mandated schemes. The absence of a ceiling, in combina- tion with limits on the pension assessment base (that is, the wages on which benefits are based, an issue discussed below) introduces progressiveness, which has implications not only for retirement savings but also for labor supply decisions for higher-income workers. BENEFIT CALCULATIONS. Benefit calculations changed greatly in many of the study countries as a result of their reforms. By construction the pen- sion assessment base in Poland (which introduced an NDC first-pillar scheme) and in Croatia, Romania, and the Slovak Republic (all of which introduced first-pillar schemes based on points) represents the revalu- ing of lifetime contributory income. The remaining four study coun- tries, which introduced only parametric reforms, also extended the pension assessment, to 18 years in Slovenia, 30 years by 2016 in the Czech Republic, and lifetime earnings in Bulgaria and Hungary. Explicit and implicit rates of accrual (that is, the percentage of the assessment base that is effectively replaced in benefits for each year of contributory service) have also been reduced, particularly in countries that intro- duced second-pillar schemes. Benefits provided by funded second-pillar defined-contribution schemes will be based on accumulated contribu- tions and investment income, net of fees and expenses, at retirement. In most cases the rules governing the payout of benefits and the institu- tional arrangements required for the payout phase have yet to be estab- lished, although most countries appear to envisage the provision of annuities by private life insurance companies.4 RETIREMENT AGES. Retirement ages are being raised in all of the study coun- tries. Retirement ages for men have been increased (or are in the process of being increased) to 62 in Hungary and the Slovak Republic; to 63 in Bulgaria, the Czech Republic, and Slovenia; and to 65 in Croatia, Poland, and Romania. Retirement ages for women are also being increased (typi- cally to 60), although they remain below those for men. Only Hungary and the Slovak Republic currently intend to establish gender neutrality. BENEFIT INDEXATION. Benefit indexation for first-pillar benefits is now automatic in all eight study countries, although the form of indexation Introduction, Summary, and Policy Conclusions 27 varies. Slovenia indexes benefits on the basis of nominal wage growth. Five countries use Swiss indexation (whereby benefits are indexed using a combination of inflation and nominal wage growth). These include Bulgaria, Croatia, and the Slovak Republic (which weight inflation and nominal wage growth equally); Poland (where inflation is given a weight of 80 percent); and the Czech Republic (where benefits are indexed on the basis of inflation plus one-third of real wage growth). Given that nominal wages tend to rise faster than prices, these differences have implications for the future adequacy of benefits and the fiscal sustainabil- ity of first-pillar schemes. Benefit adequacy tends to be negatively effected, fiscal sustainability positively. The indexation of second-pillar benefits includes price-indexed annu- ities (in Croatia and Poland) and Swiss indexation (in Hungary). It has not yet been determined in all of the study countries. Disability benefits. Disability benefits in all eight study countries con- tinue to be provided almost entirely through first-pillar arrangements, in most cases with only tenuous integration with second-pillar schemes. The enduring linkage between provisions for disability and old-age pensions may merit reconsideration, because the risks of aging (including the risk that people might outlive their savings) need not be linked, practically or theoretically, with the risks of disability and should be assessed and priced independently (Holzmann and Hinz 2005). As a result of this enduring linkage, eligibility criteria and the design of disability benefits reflect the provisions of the old defined- benefit systems (table 1.6). In all of the study countries, the vesting period for disability bene- fits increases with the age of the insured, typically reaching five years of contributory service by age 30 or (alternatively) one-third of an insured's working life from age 20 onward. Contributory service credit is awarded for service that would have been performed from the point of an individual's disability to the normal retirement age (the generos- ity of this credit varies across the study countries). Benefit eligibility requires an individual to have lost at least 30­67 percent of his or her working capacity, depending on the country. Some countries provide both partial and full disability pensions, while others provide only full disability pensions. Most countries have shifted from defining disabil- ity as the loss of capacity to perform a particular job to the loss of capacity to perform any job. Benefit determination--with regard to both the pension assessment base and the benefit rate--is based on 28 Table 1.6 Eligibility for and Benefits Provided by Disability Pensions in Eight CESE Countries Country Vesting period Contribution rate Eligibility Benefit rate Partial pension Bulgaria Under age 20: No specific At least 50 percent loss 1 percent per year -- No minimum contribution rate in working capacity Age 20­25: 1 year Age 25­30: 3 years Over age 30: 5 years Croatia Minimum coverage of No specific Permanent loss in 1 percent per year for 80 percent one-third of working contribution rate capacity for general average worker life above age 20 (26 disability pension; at for individuals with a least 50 percent loss university degree) in capacity for partial disability pension Czech Republic Under age 20: Less No specific At least 66 percent Full disability: Flat bene- Flat benefit + 0.75 per- than 1 year contribution rate loss in capacity for fit + 1.50 percent a year cent per year Age 20­22: 1 year full disability; at least Partial disability: Flat Age 22­24: 2 years 33 percent loss in benefit + 0.75 percent Age 24­26: 3 years capacity for partial a year Age 26­28: 4 years disability pension Over age 28: 5 years Hungary Under age 22: 2 years No specific At least 67 percent loss 37.5­100.00 percent of 37.5­63.00 percent of Age 22­24: 4 years contributions in capacity to work average individual average individual Age 25­29: 6 years (estimated to be earnings, depending earnings Age 30­34: 8 years roughly 4 percent) on level of disability Age 35­44: 10 years and years of service Age 45­54: 15 years Age 55 and above: 20 years Poland Under age 20: 1 year 6.00 percent (4.5 percent Total or partial Flat benefit of 75 percent of total Age 20­22: 2 years by employer, 1.5 per- incapacity to work 24.00 percent of disability pension Age 22­30: 4 years cent by employee), up reference wage + 1.30 Over age 30: 5 years to ceiling of 2.5 times percent for each average wage year of contribution + 0.07 percent for each noncontributory year Romania Under age 25: 5 years No specific contribution At least 50 percent loss Calculated on the basis Reduced benefits based Age 25­31: 8 years rate in capacity to work of number of points on level of disability Age 31­37: 11 years Age 37­43: 14 years Age 43­49: 18 years Age 49­ 55: 22 year Over age 55: 25 years Slovak Republic Under age 20: Less than 6 percent (3 percent At least 40 percent loss 1.19 percent per year Prorated if disability is 1 year by employer, 3 percent in capacity to work 40­70 percent Age 20 ­22: 1 year by employee), up to Age 22­24: 2 years ceiling of 3 times Age 24­26: 3 years average wage Age 26­28: 4 years Over age 28: 5 years Slovenia Contributed at least No specific At least 30 percent Based on level of Pro-rated on the basis of one-third of the contributions loss in capacity disability: 10­24 per- level of disability period between age to work cent of minimum pen- 20 and time of sion for full pension disability qualifying period 29 Sources: European Commission 2007, national ministries, and national social security institutions. For details, see the individual country chapters. -- = Not available. 30 Adequacy of Retirement Income after Pension Reforms old-age benefit formulas. This is also true for indexation policies. Disability benefits from second-pillar schemes generally take the form of lump-sum distributions or annuities. In some countries (such as Croatia), accumulated funds in second-pillar accounts are transferred to the accounts of the first-pillar scheme when first-pillar disability benefits exceed the combined benefit that would have been paid under both the first and second pillars. Survivor benefits. Survivor benefits in all eight study countries con- tinue to be closely linked to provisions governing old-age pensions (table 1.7). Eligibility is tied to the eligibility of the deceased for an old-age or disability pension. Eligible survivors include widows, wid- owers, children, and in some countries parents and siblings who were dependent on the deceased. The benefit rate varies but is typically about 50 percent, with supplements for additional eligible survivors. Total family benefits generally cannot exceed the benefit to which the deceased was entitled. For children eligibility generally ends when they start work, finish their studies, or reach age 25 or 26. For spouses eli- gibility depends on whether the spouse is capable of working (that is, not disabled, not caring for a child, and not too old work). If spouses are capable of working, benefits can be limited to a year (as they are in the Czech Republic, Hungary, Poland, Slovak Republic, and Slovenia). Rules regarding remarriage vary: eligibility ends only for spouses below retirement age in Hungary, and ends immediately in Bulgaria. Accumulated assets in second-pillar accounts are typically disbursed in a lump sum or installments if the deceased was of working age. If the deceased was already receiving an annuity, continuation of benefits depends on whether a single or joint annuity had been purchased. Overall, reforms left provisions governing survivorship largely intact, although in some cases eligibility criteria were tightened. No efforts have been made by the study countries to integrate survivor benefits with second-pillar provisions (Holzmann and Hinz 2005). Given that female participation in the workforce is likely to expand and divorce rates are expected to continue to rise, survivor benefits for spouses may merit reconsideration. Structure of Third- and Fourth-Pillar Schemes Both the availability of voluntary retirement savings programs to supple- ment the benefits of mandated schemes and access to health care are cen- tral to the design of a pension system. This section examines both. Table 1.7 Eligibility for and Benefits Provided by Survivor Pensions in Eight CESE Countries Orphan Orphan replacement Country Eligibility Spouse replacement rate Benefit duration Remarriage test age limit rate Total family benefit Bulgaria Eligibility of deceased One survivor: 50 percent For life if spouse is Pension ceases if 26 One survivor: 50 100 percent regard- for old-age or disabili- Two survivors: disabled or has survivor remarries percent less of number of ty pension 75 percent reached retirement Two survivors: 75 survivors, Three or more survivors : age percent Three or 100 percent or 20 per- more survivors: cent of pension of 100 percent deceased spouse as supplement to surviv- ing spouse's personal pension Croatia Eligibility of deceased 70 percent if spouse is For life unless spouse Pension ceases if 26 70 percent if orphan One survivor: 70 for old-age or disabili- only survivor remarries; also for life survivor remarries is only survivor percent of de- ty pension; minimum upon remarriage if and is younger ceased's pension of 5 years of coverage spouse remarries but is than 50 unless dis- Two survivors: 80 or 10 years of qualify- older than 50 and dis- abled percent ing periods by the de- abled Three survivors: 90 ceased percent Four or more sur- vivors: 100 percent Czech Eligibility of deceased Flat benefit + 50 percent For life if spouse is 70 Pension ceases if 26 Flat benefit + 40 No maximum Republic for old-age or deceased's pension percent disabled, tak- survivor remarries percent disability pension ing care of a child or a dependent parent, or 31 is 55 (women) or 58 (men); otherwise one year (continued) 32 Table 1.7 Eligibility for and Benefits Provided by Survivor Pensions in Eight CESE Countries (continued) Orphan Orphan replacement Country Eligibility Spouse replacement rate Benefit duration Remarriage test age limit rate Total family benefit Hungary Eligibility of deceased 50 percent of deceased's For life if spouse is dis- Remarriage test 25 30 percent; Cannot exceed for old-age or disabili- pension (20 percent if abled, caring for at before retirement 60 percent of high- benefit to which ty pension receiving own pension) least two children, or is age, none there- er of the two pen- deceased was past retirement age; after sions if orphan lost entitled 12 or 18 months if both parents spouse is caring for a child Poland Eligibility of deceased 85 percent deceased's For life if spouse is dis- Benefits paid even if 25 85 percent if sole Two survivors: 90 for old-age or pension if sole surviv- abled, taking care of a survivor remarries survivor percent disability pension ing relative child, or above age 50; Three and more otherwise one year survivors: 95 percent Romania Eligibility of deceased 50 percent of deceased's For life if married for 15 Benefits paid even if 26 50 percent if sole Two survivors: 75 for old-age or pension if married for years; married for 10 survivor remarries survivor percent disability pension 15 years; 0.5 percent for years with benefit re- Three or more sur- each month less than duction; disable and vivors: 100 percent 15 years up to a mini- married for at least one mum of 10 years; 50 year. Otherwise and percent of deceased's with children, tempo- pension if spouse is dis- rary until youngest abled and married for child reaches age 7. at least for one year; 50 Else for 6 months if percent if spouse has none of spouse children under age 7 replacement conditions is met Slovak Re- Eligibility of deceased 60 percent of For life, if spouse is Pension ceases if 26 40 percent 100 percent, public for old-age or deceased's pension 70 percent disabled, survivor remarries regardless of disability pension is caring for a child, or number of is at retirement age; survivors otherwise, for one year Slovenia Eligibility of deceased 70 percent of For life, if spouse is Benefits cease if 26 70 percent if sole 100 percent, for old-age or deceased's pension if 70 percent disabled, survivor remarries beneficiary regardless of disability pension sole beneficiary; is taking care of a before reaching number of 15 percent of de- child, or is at retire- retirement age, survivors ceased's pension if ment age; otherwise, unless incapable receiving own pension for one year of working Sources: European Commission 2007, national ministries, and social security institutions. For details, see the individual country chapters. 33 34 Adequacy of Retirement Income after Pension Reforms Voluntary old-age schemes. By 2006, all eight study countries had intro- duced voluntary private pension schemes. The primary public-policy pur- pose of these schemes was to offer individuals a credible vehicle for saving for retirement and to provide incentives for them to actually do so. To safeguard the role of these savings schemes in providing old-age income support, their products (and the providers who sell them) are typically subjected to more rigorous regulation and supervision than are nonretire- ment savings programs. The provisions governing these schemes are similar across the study countries (table 1.8). All are tax-advantaged, but most countries impose limits on preferential tax provisions. Individuals are either allowed to deduct contributions from their taxable income or provided with other sorts of subsidies. In all countries but Croatia, employers who contribute on behalf of their employees are allowed to deduct contributions from income subjected to enterprise taxation. In all countries but Bulgaria and Croatia, schemes are subjected to vesting requirements (minimum partici- pation periods). In all countries but Hungary, individuals must reach a minimum age to access the funds in their accounts. None of the countries imposes special rules on disbursement; all allow funds to be withdrawn as a lump sum. Health care provisions for retirees. Access to health care is critically important to the issue of pension adequacy, because (in the absence of affordable health care) many elderly people would be forced to spend a main share of their retirement income on private health insurance, if one assumes such coverage were even available. The adequacy of pension ben- efits is affected by official out-of-pocket costs (including copayments for doctor visits and pharmaceuticals) and unofficial costs (including the enduring tradition from the communist era of paying tips to staff mem- bers of the health care providers). All study countries provide mandatory health insurance for their active populations, including the self-employed (table 1.9). Health insurance schemes are financed by contributions from employers and employees. Benefits include cash payments (for example, sick pay), in-kind health care services, pharmaceuticals, and other items. Elderly people receiving a public pension (and their dependents) have access to the same benefits as the active population, but they typically pay a lower contribution rate (part or all of the difference is paid by the government or by the pension authority on their behalf).5 The share of private expenditures in total health care expenditures in the study countries falls roughly in line with Table 1.8 Voluntary Pension Provisions in Eight CESE Countries Tax advantages Contributions tax Lump-sum payments Country Vesting period Retirement age to participants deductible by employers possible in retirement Bulgaria No 58 (men), 55 (women) Yes Yes Yes Croatia Noa 50 Yes No Yes Czech Republic 5 years 60 Yes Yes Yes Hungary 10 years No set retirement ageb Yes Yes Yes Poland 5 years 60 Yes Yes Yes Romania 90 months 60 Yes Yes Unknownc Slovak Republic 10 years 55 Yes Yes Yes Slovenia 10 years 58 Yes Yes Yes Sources: European Commission 2007, national ministries, and national social security institutions. For details, see individual country chapters. a. There is no specific vesting period. Benefits can be collected once an individual retires from the mandated schemes or upon reaching age 50. b. Benefits can be withdrawn after 10 years. c. Regulations related to the payout phase have not yet been issued. 35 36 Table 1.9 Health Care Provisions for Contributors and Retirees in Eight CESE Countries Public and private expenditure on Contribution rate Out-of-pocket health as percent (percent) Total 2005 expenditure as of total health health expendi- percent of pri- Active population expenditure ture as percent vate expendi- Country Employer Employee Total Retirees of GDP Public Private ture on health Bulgaria 3.0 3.0 6.0 None 7.7 60.6 39.4 96.3 Croatia 15.0 0.0 15.0 None 7.4 81.3 18.7 93.6 Czech Republic 9.0 4.5 13.5 13.5 percent of state- 7.1 88.6 11.4 95.3 defined wage level, paid by the government Hungary 4.0 11.0 15.0 None 7.8 70.8 29.2 86.8 Poland 0.0 9.0 9.0 9.0 percent of pension 6.2 69.3 30.7 85.1 Romania 7.0 7.0 14.0 7.0 percent of pension 5.5 70.3 29.7 85.0 Slovak Republic 10.0 4.0 14.0 14 percent of minimum 7.0 74.4 25.6 88.1 wage, paid by the government Slovenia 6.56 6.36 12.92 5.0 percent of pension 8.5 72.4 27.6 45.0 Source: European Commission 2007, national ministries, and national social security institutions. For details, see individual country chapters. Note: All eight countries provide public health insurance. Introduction, Summary, and Policy Conclusions 37 the shares observed in Organisation for Economic Co-operation and Development (OECD) countries, but--in contrast to OECD countries-- private expenditures are related largely to out-of-pocket costs, not to the cost of private health insurance. Assessment of the Performance of Pension Systems The primary objectives of the eight country studies was to assess the per- formance of their reformed pension systems in a steady state (that is, as if their reforms had been in operation for long enough that current workers had always been subjected to the new rules). Because such an evaluation cannot be conducted ex post for many years, analytical tools had to be designed and applied specifically for this purpose, and benchmarks had to be developed against which performance could be evaluated. Given that earnings trajectories vary across countries (as a result of which, average benefits can vary across countries even if their pension systems rely on exactly the same provisions and average earnings are the same), the eval- uation also required the development of a methodology for assessing countries both individually and collectively in a way that permitted com- parisons across the sample. The World Bank has established four principles for evaluating public pension systems, which together should guide the process of pension reform (see Holzmann and Hinz 2005). Broadly speaking, these princi- ples include the adequacy and security of benefits, the affordability of contributions, the sustainability of the system over time, and the robust- ness of the system in the face of demographic changes and macroeco- nomic shocks. This section focuses primarily on the adequacy of benefits and the financial sustainability of the first- and second-pillar earnings- related pension schemes. The remaining principles are mentioned only briefly. Adequacy is analyzed through the lens of net replacement rates. Financial sustainability is evaluated using projections of pension expen- ditures and revenues. Benefit Adequacy Replacement rates are a useful yardstick for measuring the adequacy of pension benefits, because they express benefits relative to preretirement earnings, thereby indicating the degree to which income is replaced when workers retire. Two variants are commonly used. Gross replace- ment rates compute income replacement as the ratio of gross benefits paid to pretax preretirement earnings. Net replacement rates compute 38 Adequacy of Retirement Income after Pension Reforms income replacement as the ratio of benefits received (that is, after the payment of taxes and other levies, including contributions for social insur- ance) to posttax preretirement earnings. In general, net replacement rates are a more useful measure of benefit adequacy, because they capture the degree to which actual take-home pay is replaced when workers retire. Replacement rates are a function of the formula governing pension benefits; an individual's contribution history; and, in the case of net replacement rates, the rules of income tax, social security contributions, and other relevant levies. The benefit formula establishes the degree to which the system redistributes income across individuals of different levels of preretirement earnings. Progressive systems provide higher levels of income replacement to people with lower levels of preretirement income. In general, the degree to which a system is redistributive depends on (a) the existence (and value) of flat transfers and minimum pension guarantees, (b) the degree to which benefits are earnings related, and (c) the existence of ceilings on earnings subject to contributions. An individual's contribution history can be characterized by his or her age of entry into the labor force, contribution density, and decisions regarding the timing of retirement. To some degree, these three factors are influenced by the incentives embodied in the pension system. The tax and contribution sys- tem influences net replacement rates through the progressiveness of the income tax formula, which taxes (higher) income during a worker's active life more than it taxes (lower) pension benefits in retirement. In addition, social security levies (for pensions; unemployment; health care; and, at times, housing and family benefits) are typically reduced or elim- inated altogether in retirement. These benefits are particularly important for low- to middle-income groups. Benchmarks need to be established to evaluate the adequacy of the income replacement provided by the earnings-related pension schemes. Unfortunately, there is no consensus on what constitutes adequacy. According to one widely respected definition, pensions are adequate when they are sufficient to prevent poverty among the elderly and to provide the vast majority of people with a reliable mechanism for smoothing income over their lifetime. Even with a definition, however, establishing bench- marks is problematic, because attitudes vary across countries as a result of social and cultural perceptions. Moreover, benchmarks ignore the other factors that affect the welfare of the elderly--and that also vary across countries--including the existence and generosity of health insurance and long-term care, the cost of housing, the structure of traditional living arrangements, the presence of informal intrafamily or intergenerational Introduction, Summary, and Policy Conclusions 39 sources of financial and nonfinancial support, and the availability and security of other mechanisms for people to save for their own retirement. One reputable nine-country study (OECD 2001) observes that living standards are roughly comparable for people 10 years older than the nor- mal retirement age and people 15 years younger than the normal retire- ment age when retirees have disposable income equal to roughly 80 percent of the disposable income of working-age people. In part, this is attributable to the fact that retirees have no work-related expenses (they do not have to commute or buy special clothing or uniforms, for exam- ple). This finding, however, does not imply that mandatory first-pillar pension schemes should actually target an 80 percent net replacement rate. To the contrary, in middle- and high-income countries, one can rea- sonably expect individuals to save for their own retirement--and the empirical evidence suggests that, in practice, they do so.6 There is also some evidence to suggest that the ratio between pre- and postretirement income is somewhat independent of the income replacement mandate of the public pension system. Put simply, individuals tend to save more in countries with more modest mandates (and vice versa). Because mechanisms for saving--such as voluntary pension schemes­ exist in all of the study countries, it would seem reasonable to expect middle- and higher-income workers to save enough to finance at least 25 percent, if not closer to 50 percent, of this 80 percent income replace- ment target. Given this, three benchmarks are provided: a 40 percent net replacement rate (which implies that individuals would be expected to save enough to finance half of the total income replacement target); a 60 percent net replacement rate (which implies that individuals would be expected to finance a quarter of the target); and an 80 percent net replace- ment rate (which implies that individuals, most of whom would be low- income earners, would not be expected to contribute anything toward the target).7 In the analysis conducted for each of the eight country-specific studies, the results of which are summarized below, adequacy is evaluated against these three benchmarks, as well as in conjunction with the average net replacement rate observed in 53 countries; the average net replace- ment rate observed for selected countries in Europe and Central Asia; and the average poverty line for the reviewed countries.8 In order to estimate gross and net replacement rates, we use the APEX model to consider two critical dimensions: earnings levels and contribu- tion periods.9 This model generates estimates for replacement rates under steady-state assumptions (that is, as if the rules of the reformed pension scheme had been in place over the entire active life of the individual). 40 Adequacy of Retirement Income after Pension Reforms Because life expectancies at retirement are projected to increase over time--which will affect the benefits paid by defined-contribution pen- sion schemes--a reference year must be chosen. For this study, 2040 is used, because it provides a sufficiently long contribution period over which to approximate steady-state conditions. Gross and net replacement rates are considered both across incomes (where the income spectrum is expressed relative to average earnings, ranging from half the average to twice the average) and as a function of patterns of contributory service (which are captured by the age at which someone is assumed to enter the workforce, the degree to which he or she works continuously or with interrupted service, and the age at which he or she elects to leave the workforce permanently). Full-career workers are considered first, fol- lowed by partial-career workers (that is, people with intermittent pat- terns of formal-sector employment). Income replacement for full-career workers. For this analysis, a full career is defined as continuous employment from age 20 to the retirement age in effect once reforms have been fully implemented. For men, this age ranges from 62 to 65 across the eight study countries. GROSS REPLACEMENT RATES. Gross replacement rates were estimated using total benefits from all mandated pension provisions (that is, sec- ond-pillar benefits are included in the analysis for countries with second- pillar schemes).10 Figure 1.3 shows gross replacement rates as a function of preretire- ment income relative to the economywide average earnings. Analysis of figure 1.3 yields the following observations: · Gross replacement rates differ substantially across the study coun- tries--and the variation widens as earnings increase. At half the aver- age earnings, gross replacement rates range from 50.00 percent (in Croatia) to 74.8 percent (in Bulgaria). At twice the average earnings, gross replacement rates range from 30.1 percent (in Czech Republic) to 74.8 percent (in Bulgaria). · Countries have introduced reforms that impose greater actuarial neu- trality (that is, NDC schemes, point systems, and similar parametric reforms). Bulgaria, Romania, Poland and the Slovak Republic provide gross replacement rates that are flat across the income spectrum but at different levels. Adding a second pillar does not change this result, Introduction, Summary, and Policy Conclusions 41 Figure 1.3 Gross Replacement Rates for Male Full-Career Workers in Eight CESE Countries 90 high benchmark 80 70 ent)c 60 (per 50 estar 40 ement 30 eplacr 20 middle benchmark 10 low benchmark 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage Bulgaria Romania Poland Slovak Republic Hungary Slovenia Croatia Czech Republic Source: World Bank staff estimates. because the benefits paid by defined-contribution schemes are, by construction, directly determined by contributions. · As a result of flat benefit provisions intended to affect income redistri- bution, gross replacement rates fall as income rises in Croatia (which has a two-tiered first-pillar point scheme) and the Czech Republic (which has a two-tiered first-pillar defined-benefit scheme). Slovenia's pension scheme is also progressive, but only at higher income levels. The Hungarian system has recently become progressive, as a result of the (recently introduced) net income base used to calculate benefits. Gross replacement rates for countries with second-pillar schemes are time dependent, because second-pillar benefits are computed as annuities, which vary based on conditional life expectancy. Because life expectancy is increasing (at the rate of roughly two years per decade), the value of 42 Adequacy of Retirement Income after Pension Reforms annuities will fall accordingly. if one assumes that funded schemes provide roughly a third of total gross income replacement, an increase in life expectancy of 10 percent per decade reduces total gross replacement rates by roughly 2 percentage points for individuals of the same age who retire 10 years apart. A similar reduction in gross levels of income replacement can occur in first-pillar schemes. Indeed, one of the objectives of an NDC scheme (such as the one introduced in Poland) is to reduce the vulnerabil- ity of pension systems to the uncertainties surrounding future changes in life expectancy. When the impact of rising life expectancy on both pillars is considered, an increase in life expectancy of 10 percent per decade reduces total gross replacement rates by some 6 percentage points over 10 years. In Croatia, such a reduction is affected through the use of Swiss reval- uation of the points an individual accumulates while working. However, because the flat benefit component of its first-pillar scheme is, by defini- tion, revalued to be 25 percent of the average wage, rising life expectancy will have a smaller impact on gross replacement rates--less than 1 per- centage point for the first pillar alone and roughly 3 percentage points for both pillars combined over a 10-year-period. NET REPLACEMENT RATES. Gross replacement rates are of limited value in assessing the adequacy of retirement income. Net replacement rates are a better indicator. Figure 1.4 shows the net replacement rates that result from the gross replacement rates shown in figure 1.3 after the application of country-specific rules regarding income taxes and social security con- tributions. Several observations emerge from analysis of figure 1.4: · Net replacement rates are well above their corresponding levels of gross income replacement in all eight study countries. The Slovak Re- public and the Czech Republic--with net replacement rates of 66 percent and 102 percent, respectively--define the range at half aver- age earnings. At twice average earnings, the Czech Republic (42 per- cent) and Hungary (100 percent) define the range. Rising replacement rates in countries such as the Slovak Republic are due entirely to their progressive income tax rates, not to structural characteristics of their pension schemes. · Net replacement rates for most countries remain broadly constant in the range of 60­100 percent, with some country-specific variation. In Croatia and the Czech Republic, net replacement rates fall substantially Introduction, Summary, and Policy Conclusions 43 Figure 1.4 Net Replacement Rates for Male Full-Career Workers in Eight CESE Countries 120 100 80 (percent) rate 60 40 replacement 20 low, middle and high benchmarks 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage Hungary Romania Bulgaria Poland Slovak Republic Slovenia Croatia Czech Republic ECA Average World Average Average Poverty Line Source: World Bank staff estimates. across the income spectrum. In Hungary and the Slovak Republic, net rates increase across the income spectrum. In Bulgaria and Poland net rates remain essentially flat. · All of the study countries provide levels of net income replacement for full-career male workers that are higher than the middle benchmark (60 percent); in several countries (Bulgaria, Hungary, Romania and Slove- nia), net replacement rates are close to or above the high benchmark (80 percent). Only in the Czech Republic at twice average earnings do net replacement rates fall near the low benchmark (40 percent). Income replacement for partial-career workers. Estimating replacement rates for full-career workers under the earnings-related pension schemes 44 Adequacy of Retirement Income after Pension Reforms of the eight study countries is a useful undertaking, because it establishes an upper benchmark for what these systems are promising to deliver. Not everyone works from age 20 to the statutory retirement age, however. Many individuals enter and exit the labor force (often at different ages and for different periods of time) and earn different wages while working (figure 1.5).11 To study the adequacy of benefits for partial career workers, we exam- ine three stylized cases. These cases include career type A (someone entering the labor force at age 25 who works continuously for a period of years before leaving the workforce at some point between the ages of 50 and 70 and then claims a benefit); career type B (identical to career type A, except that the worker enters the workforce at age 30 and leaves no earlier than age 55); and career type C (identical to career type A, except that the individual contributes in only three years out of four while in the labor force). In cases where the withdrawal from the formal labor market occurs before the statutory retirement age, the pension is claimed (and the replacement rate calculated) only at the later age. For withdrawals after the statutory retirement age, the ages coincide. Net replacement rates for these stylized, partial-career, middle-income workers are shown in figure 1.5. Several observations emerge from analysis of figure 1.5: · Exiting the labor market at an early age exacts a significant cost in terms of net replacement rates. Exiting the labor market at age 45 and then drawing a pension upon reaching the minimum retirement age reduces levels of net income replacement by one-third or more compared to leaving the labor market at age 70 and drawing a pension immediately. Despite this reduction, net replacement rates remain well above coun- try-specific poverty levels, even at one-half of the average earnings. · Several of the study countries reward workers who elect to remain in the labor market after reaching the standard retirement age (62­65 for men). In these countries, each year of additional contributory serv- ice beyond the normal retirement age is rewarded with even more pronounced increases in net replacement rates. In Slovenia, however, net replacement rates are flat at higher retirement ages, which does nothing to encourage older workers to remain in the workforce. · Entering the labor market later, at age 30 instead of age 25, exacts a significant cost in terms of net replacement rates. For workers retiring Figure 1.5 Net Replacement Rates for Male Partial-Career Workers in Eight CESE Countries Bulgaria 140 Croatia 140 Czech Republic 140 Hungary 140 120 120 120 120 100 (percent) 100 (percent) 100 (percent) 100 (percent) 80 rates 80 rates 80 rates 80 rates 60 60 60 60 40 40 40 40 20 replacement 20 replacement 20 replacement 20 replacement net net net net 0 0 0 0 45 50 55 60 65 70 45 50 55 60 65 70 45 50 55 60 65 70 45 50 55 60 65 70 exit age from labor market exit age from labor market exit age from labor market exit age from labor market Poland Romania Slovak Republic Slovenia 140 140 140 140 120 120 120 120 100 (percent) 100 (percent) 100 (percent) 100 (percent) 80 rates 80 rates 80 rates 80 rates 60 60 60 60 40 40 40 40 20 replacement 20 replacement 20 replacement 20 replacement net net net net 0 0 0 0 45 50 55 60 65 70 45 50 55 60 65 70 45 50 55 60 65 70 45 50 55 60 65 70 exit age from labor market exit age from labor market exit age from labor market exit age from labor market Type A Type B Type C Entry age: 25 Entry age: 30 Entry age: 25 Contribution density: 100% Contribution density: 100% Contribution density: 75% 45 Source: Authors' estimates based on European Commission 2007, national ministries, and national social security institutions. For details, see individual country chapters. Note: The four dark lines adjacent to the y-axis represent the 80 percent, 60 percent, and 40 percent benchmark; the poverty line is expressed as a percentage of the average wage. See text for description of career types. 46 Adequacy of Retirement Income after Pension Reforms at age 65, entering the labor market at age 30 results in levels of net income replacement that are lower by 5­10 percentage points in most countries. The cost of entering the labor market later are compounded for those who also exit the labor market early. · Working intermittently is costly in terms of net income replacement. Someone who enters the workforce at the same age but who contributes for only three years out of four will receive a net replacement rate 8­20 percentage points lower than someone who contributes continuously. In most of the study countries, career type A and B workers will attain the lowest of the three benchmarks before reaching retirement age, while career type C workers must work until retirement age. Career type A and B workers can often attain the middle benchmark provided they work beyond retirement age. In many cases, career type C workers will not attain the 80 percent benchmark even if they work until 70. Benefit indexation. The adequacy of pension benefits is determined not only by the level of an individual's replacement rate at retirement but also by how the benefits are adjusted over retirement in response to changing prices or overall living standards. The debate among pension economists and practitioners continues over whether the optimal path of retirement consumption for a fixed level of pension wealth should rise or fall and, hence, a higher initial replacement rate combined with lower indexation or a lower replacement rate combined with higher indexation. But most would agree that benefits from mandated schemes should at least main- tain their real value (that is, be indexed to prices). In a growing economy in which wages are rising faster than inflation, however, price indexation results in a continuous softening of the relative consumption position of retirees over time. The practical consequence of price indexation is that the benefits of two otherwise identical workers will differ as a function of how long they have been retired. In situations in which real annual wage growth is high (say, 2 percent or more), the resulting differences may become substantial. For this reason, a number of countries--in both CESE and around the world--index benefits using the rates of growth of both prices and wages, in varying proportions. Seven of the eight study countries provide for the indexation of benefits from both the first and (for countries that have them) the second pillar (the exception is Romania, which has yet to establish a policy governing the indexation of second-pillar benefits) (table 1.5). The mechanism by which Introduction, Summary, and Policy Conclusions 47 benefits are indexed, however, varies. The generosity of indexation policies is framed by two countries--Poland and Slovenia--with the latter having only a first-pillar scheme. Poland indexes benefits using 80% price and 20% wage increase, while Slovenia uses wages only. A second country without a second pillar (the Czech Republic) indexes benefits on the basis of prices plus one-third of real wage growth. Countries with both a first and a sec- ond pillar typically use Swiss indexation (half prices, half wages) for first- pillar benefits and price indexation for second-pillar benefits. The effects of indexation policies on replacement rates over retire- ment are shown in figure 1.6. To facilitate comparisons, replacement rates are normalized to 100 percent for all countries. Underlying modeling Figure 1.6 Impact of Indexation on Income Replacement (Active Earnings Units) in Eight CESE Countries 1.1 1.0 (retirement=100) 0.9 units 0.8 earnings active 0.7 in rate 0.6 replacement 0.5 of value 0.4 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 years in retirement Hungary Romania Bulgaria Poland Slovak Republic Slovenia Croatia Czech Republic Source: Author estimates. 48 Adequacy of Retirement Income after Pension Reforms assumptions--including the assumption that inflation averages 2.5 percent a year and real wage growth averages 2 percent per year--are unchanged. Changes in replacement rates are measured against full wage indexation, which is the functional equivalent of comparing postretirement benefits to the average earnings of current workers (that is, expressing them in active earnings units).12 Several observations emerge from analysis of figure 1.6: · Despite a modest assumption about the growth of real wages, the impact of less than full wage indexation on relative pension benefits in retirement is noticable in all countries except Slovenia (where bene- fits are wage indexed). · In Poland, where benefits are predominantly indexed to prices, the rel- ative income of someone retired for 10 years will be 16 percent lower than that of new retirees with the same relative preretirement wages. After 35 years (that is, for those few people living to almost age 100), relative income will be almost half that of new retirees with the same preretirement wages. · The use of both prices and wages to index first-pillar benefits in com- bination with the price indexation of second-pillar benefits--as is done in Croatia and the Slovak Republic--somewhat dampens, but does not eliminate, these changes in relative income position over time. Financial Sustainability The sustainability of pay-as-you-go public pension schemes is best evalu- ated in actuarial terms by estimating the actuarial deficit as the difference between a scheme's liabilities and assets. If the actuarial deficit is positive and large, a scheme is financially unsustainable and will require policy remediation to increase revenues or reduce expenditures. A good proxy for the actuarial deficit is the difference between present value of expected future revenues (that is, contributions and other sources of income) and expected future expenditures (that is, benefit payments, administrative costs, and other expenses) over a long projection period. The difference between the net present value of these two projected cash flows repre- sents an unfunded liability (sometimes referred to as a financing gap) on the public-sector balance sheet. Given that the eight individual country studies were also concerned with the path of revenues and expenditures Introduction, Summary, and Policy Conclusions 49 (and the resulting net cash flow deficit) over a projection period extend- ing to 2050 (figure 1.7), this more pragmatic approach to measuring sus- tainability has been used. Although the projections discussed as follows are all based on different sources (and, thus, lack strict comparability), the methodologies underlying their preparation are reasonably close and, consequently, provide a reasonable basis for comparison.13 Against this benchmark, the pension reforms of all eight study coun- tries made major progress toward addressing underlying issues of sustain- ability. Progress across the sample is uneven, however, and further reforms are needed in a number of countries. Several observations emerge from analysis of figure 1.7: · The path of projected net cash flows varies widely across the study countries. The pension systems of two counties (Croatia before 2050 and Poland thereafter) are expected to eventually reach fiscal balance Other countries (the Czech Republic, the Slovak Republic, and Slovenia) start out closer to fiscal balance and show initial improvements before deteriorating over the long term. Still others (Bulgaria, Hungary, and particularly Romania) show imbalances throughout the projection period. Given that all of the study countries are confronting broadly similar demographic challenges, differences in cash flow projections result primarily from differences in pension system design and (in countries with second pillars) varying capacities for financing the tran- sition costs associated with the introduction of funding. · Countries that introduced systemic reforms (such as Poland and Croatia) seem to have been more successful in raising retirement ages (to 65 for men and 60 for women) and in lowering levels of gross income replacement (to 61 percent and 49 percent, respectively, for someone with average earnings) than were countries with less ambitious reforms (such as Bulgaria, where retirement ages are 63 for men and 60 for women and gross replacement rates are 66 percent; Hungary, where retirement ages are 62 for both men and women and gross replace- ment rates are 77 percent; and Romania, where retirement ages are 65 for men and 60 for women and gross replacement rates are 62 per- cent). Similar design differences also appear to be driving the projected deterioration of fiscal balances in the Slovak Republic (where retire- ment ages are 62 for men and women and gross replacement rates are 57 percent) and Slovenia (where retirement ages are 63 for men and 61 for women and gross replacement rates are 62 percent). 50 Figure 1.7 Projected Pension System Fiscal Balances after Reform in Eight CESE Countries 20 Bulgaria 20 Croatia 20 Czech Republic 20 Hungary 15 15 15 15 expenditure expenditure expenditure GDP 10 GDP 10 GDP 10 GDP 10 expenditure of of revenue of revenue of 5 5 5 5 revenue revenue 0 0 0 0 percentage percentage percentage percentage 2000 2005 2010 2015 2040 2050 2005 2010 2015 2020 2025 2030 2035 2040 2050 2009 2014 2019 2024 2029 2034 2039 2044 2049 2050 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 ­5 ­5 ­5 ­5 deficit deficit deficit deficit ­10 ­10 ­10 ­10 20 Poland 20 Romania 20 Slovak Republic 20 Slovenia expenditure 15 15 15 15 expenditure expenditure GDP 10 GDP 10 GDP 10 GDP 10 expenditure revenue of revenue of of of 5 5 5 5 revenue revenue 0 0 0 0 percentage percentage percentage percentage ­5 2000 2010 2020 2030 2040 2050 ­5 2008 2010 2020 2030 2040 2050 ­5 2005 2015 2025 2035 2045 2050 ­5 2005 2010 2020 2030 2040 2050 deficit deficit deficit deficit ­10 ­10 ­10 ­10 Source: European Commission 2007, national ministries, and national social security institutions. For details, see individual country chapters. Introduction, Summary, and Policy Conclusions 51 The tools for addressing the sustained pension deficits observed in the projections for some of the study countries are limited. In theory, rev- enues can be increased and benefits reduced or delayed. In practice, the options available to policy makers are more constrained: · Pension scheme revenues can be increased by raising contribution rates, but because raising contribution rates can threaten competitive- ness--and will likely strengthen incentives for tax evasion, which is already of concern in the region--it is typically not embraced by pol- icy makers. Moreover, it would represent a reversal of policy in most of the study countries. · For countries in which a large portion of the deficit is attributable to the transition costs of newly introduced funded second pillars, policy makers might consider financing all or part of the transition costs with general revenues. This effectively represents the partial repayment of the implicit debts accrued by their first-pillar pension schemes. More- over, general revenue financing may be perceived as being more equi- table and less distortional. · Further increasing the retirement age (while concurrently adjusting benefit provisions to fix the level of income replacement awarded at retirement) is an attractive option, because it should improve both revenues and expenditures. Furthermore, raising retirement ages is the most logical and consistent policy for addressing increases in life expectancy. Actually increasing the preponderance of elderly workers in the labor force, however, may require more than simply raising statutory retirement ages and may depend on broader labor and finan- cial market reforms.14 · Alternatively--or in addition to, because the options are not mutually exclusive--expenditures can be reduced by cutting benefits. Addressing financial sustainability through benefit cuts has, of course, a direct bearing on benefit adequacy. Achieving fiscal balance by 2050 will re- quire proportionate benefit cuts--evaluated at average earnings--of 20 percent (equivalent to a 10­12 percentage point reduction in net replacement rates) in Bulgaria, the Czech Republic, and Hungary and 35 percent (equivalent to a 25­30 percentage point reduction in net replacement rates) in Romania and Slovenia. 52 Adequacy of Retirement Income after Pension Reforms · To offset reductions in first-pillar benefits, in lieu of raising the retire- ment age, governments could encourage individuals to engage in volun- tary savings. To increase income replacement by 1 percentage point, for example, a full-career worker would need to save almost 0.5 percent of his or her earnings from the age of 40 to the current age of retire- ment.15 Making up for a 10 percentage point drop in net replacement rates would, therefore, require additional savings of some 5 percentage points of earnings. Conclusions All eight of the study countries have made major progress in reforming their pension systems over the past decade. Financial sustainability--a key concern and the dominant driver behind most reforms--has been dra- matically improved in most countries. The connection between contribu- tions and benefits has been strengthened, and overall system designs are now better aligned with (and suited for) a market economy given the introduction of voluntary third-pillar schemes in all eight countries and the introduction of mandatory second-pillar schemes in six of them. A major social policy concern--and a predominant focus of this report and its underlying country studies--has been the (at times, express) concern that reforms may have exacted a high cost in terms of the adequacy of pension benefits of future retirees. In evaluating this concern, the country studies rely on an analytical toolkit that generates estimates for gross and net replacement rates under earnings-related schemes for workers with different earnings levels and lifetime patterns of contributions. These calculations were performed under steady-state assumptions (that is, as if the reforms had been in operation for long enough that current workers had been subjected to the new rules for most of their working lives). A reference year of and mortality projec- tions for 2040 were used to provide a sufficiently long contribution period to approximate steady-state conditions. Several findings emerge from these calculations: · Estimated net replacement rates (the relevant welfare indicator for consumption smoothing and the best measure of the degree to which preretirement take-home pay is replaced by disposable income in retirement) suggest that levels of income replacement for full-career workers are generally in line with regional averages and international benchmarks. In five countries (Bulgaria, Croatia, the Czech Republic, Introduction, Summary, and Policy Conclusions 53 Poland, and the Slovak Republic), net replacement rates evaluated at low and average earnings are between the middle (60 percent) and high (80 percent) benchmarks. In three countries (Romania, Hungary, and Slovenia), net replacement rates exceed 80 percent. · Estimated net replacement rates for partial-career workers--those who enter the workforce later in life, have intermittent patterns of formal-sector employment, or leave the workforce before reaching the standard retirement age--illustrate the importance of continued formal labor market participation. Although net replacement rates for most partial-career workers are well above the poverty line in most of the study countries, individuals in some countries are not ensured of achieving even the low (40 percent) benchmark. Promoting more continuous and longer labor force participation, however, is a cross- sector labor market challenge that should be addressed with macroeco- nomic and microeconomic policies that go beyond the use of pension policy to achieve adequate pensions for such workers. · All of the study countries have introduced voluntary pension schemes to provide a vehicle with which workers can supplement the benefits provided by mandated schemes. Voluntary schemes are particularly important for middle- and high-income earners, because public schemes cannot be expected to provide them with all of their income replacement in retirement. · All of the study countries have noncontributory, zero-pillar schemes to alleviate poverty among the elderly. In five countries (Croatia, the Czech Republic, Poland, Romania, and the Slovak Republic), the zero pillar is part of a broader scheme of social assistance available to every- one, regardless of age, that is intended to guarantee a minimum income level. Three countries (Bulgaria, Hungary, and Slovenia) provide an age- related social pension specifically for the elderly. Such provisions may gain importance in the future as a way of addressing income gaps and supporting those most seriously affected by the economic transition. Although progress toward addressing issues of fiscal sustainability has been made in all of the study countries, only a few are fully prepared for the inevitability of population aging. Progress remains uneven. The options available to policy makers are limited: revenues can be increased, benefits can be reduced, or retirement ages can be raised. Increasing revenues by 54 Adequacy of Retirement Income after Pension Reforms promoting formal-sector employment across all age groups or improving compliance certainly helps improve cash flows in the short-term. Such policies, however, are not a panacea for actuarially unsustainable schemes. On the contrary, an increase of employment in a pension scheme that promises benefits in excess of contributions indexed by the sustainable (implicit) rate of return creates only temporary cash flow surpluses. Because the increase in pension liabilities exceeds the increase in (pay-as-you-go) assets, the sustainability of the scheme actually dete- riorates. For this reason, promoting formal-sector employment--an objective that can and should be pursued across the region to increase benefit coverage--is no substitute for pension reforms intended to address underlying design issues of sustainability. The introduction of second-pillar pension schemes by some study countries increased, rather than reduced, fiscal pressure, because the introduction of funding means that a portion of contributions that were once available for the payment of benefits are diverted to funded accounts. Addressing fiscal imbalances exclusively through benefit cuts threatens benefit adequacy and could undermine political support for reforms. One possible alternative would be to finance transition costs using general revenues, which may have more equitable incidence and be less distortional. CESE countries have been reluctant to use budgetary financing for this purpose given the Maastricht fiscal criteria and their desire to join the euro area. Given that the rules governing the accounts of public pension systems are scheduled to be revised in the European System of National Accounts to reflect the actual debt-reducing nature of transition deficits, CESE countries may have more room in the future to reconsider this option. Achieving fiscal balance by 2050 for unfunded first-pillar schemes in some of the study countries on the basis of benefit cuts alone would require that average replacement rates be reduced by 35 or more percent- age points. In addition to not being politically feasible, such drastic cuts would likely jeopardize the adequacy of pension benefits. An alternative approach (which could be combined with more modest cuts in benefits or implemented on its own) would be to move toward fully price index- ing benefits in disbursement. Although the preservation of purchasing power across retirement (for benefits that were adequate in terms of income replacement at retirement) is consistent with adequacy consider- ations, it would result in a reduction in purchasing power relative to the active population and younger retirees. This may not be politically viable in periods of high real-wage growth. Thus, the introduction of full price Introduction, Summary, and Policy Conclusions 55 indexation may require additional and discretionary increases in benefits conditioned on developments with real wages and budgetary resources. Given a lack of viable alternative policy options, the challenge of pop- ulation aging in all countries in the region demands that more decisive steps be taken to encourage and enable extended labor force participation among the elderly. Contribution rates for pensions and other social pro- grams are already very high and may be partly to blame for the existence of sizable informal labor markets in much of the region. Further benefit cuts may unduly undermine benefit adequacy in some countries. Although all of the study countries raised retirement ages as part of their reforms, legal and actual retirement ages are still generally low in compar- ison to most OECD countries and in comparison to past and future gains in life expectancy at retirement. Moreover, all of the study countries except Hungary continue to allow women to retire substantially earlier than men, despite women's having substantially higher life expectancies. Although benefit adequacy seems broadly ensured for most workers in all of the study countries--particularly if retirement ages are raised (and equalized) in line with life expectancies and if policy changes regarding the financing of transition costs are enacted--CESE countries may need to think about temporary measures to provide income support for the lost generation that is now emerging from the transition. Many members of this generation suffered from low earnings and patchy formal- sector employment in the years following the end of central planning; they now risk receiving very low pension benefits, if they qualify at all (Augusztinovics and Köllő 2009). Although addressing their needs by making benefits more generous may be tempting (and politically expedi- ent), a longer-term perspective would suggest that targeted transitional measures may be more effective without undermining progress toward fis- cal sustainability. Most of the study countries have introduced painful reforms that will bear fruit in the future, weakening, one hopes, some of the incentives driving labor market informality. Permanent changes to meet the special needs of this lost generation would not only undermine sustainability (and adequacy, because unsustainable schemes threaten the adequacy of benefits for future generations of retirees), but also weaken the currently tight linkages between contributions and benefits and encourage continued labor market informality in the future. Last but not least, the projections for second-pillar pension schemes discussed earlier assumed that the schemes would earn net rates of return 1.5 percentage points higher than earnings growth. Although such returns are in line with historical performance in developed countries and are 56 Adequacy of Retirement Income after Pension Reforms substantially lower than historical average performance in emerging markets (Musalem and Bebczuk, 2009, forthcoming), there is no assur- ance that the funded schemes of the study countries will achieve this benchmark. And the current financial crisis has provided a painful reminder that these long-term target rates can be put into jeopardy by exceptional events. But even before the crisis, the performance of pen- sion funds in the region had been highly uneven and often disappoint- ing. Improving their performance calls for a review of pension fund structures and accelerated financial sector reforms if these schemes are to live up to the return expectations of future retirees in light of popu- lation aging (Holzmann, 2009). Notes 1. Although the World Bank undoubtedly influenced the thinking of policy mak- ers through its analytical work, access to information, and capacity-building measures, it has never imposed such an approach to pension reform, as has sometimes been suggested (Orenstein 2008). 2. Benefits from noncontributory schemes are not necessarily tax exempt. Taxing them as ordinary income (as is done in Australia and New Zealand, where benefits are not asset tested) provides a way of clawing back benefits from income-rich retirees. 3. See Tesliuc and others (2008) for a review of the targeting efficiency of social assistance schemes in countries in Europe and Central Asia. 4. See Rudolph and Rocha (2008) for a discussion of the status of and lessons learned from the preparations for the payout phase in Chile. 5. A lower contribution rate for the elderly--or the payment of contributions on their behalf--does not necessarily imply that the elderly are receiving a sub- sidy. From a life-cycle perspective, what matters is lifetime contributions rel- ative to lifetime benefits. Higher contribution rates for people of working age can compensate for lower contribution rates for when those people are no longer working. The fact that contributions are income related, however, typ- ically effects major redistribution from people who are comparatively wealth- ier to those who are comparatively poor. 6. In Chile, for instance, 70 percent of retirees from the mandatory public pension system own their home, which is a form of savings (see Valdčs-Prieto 2008). 7. These benchmarks approximate the standards developed by the International Labour Organization (ILO) and the Council of Europe (1990). ILO Convention 102 of 1952 sets a minimum benefit equal to 40 percent of the reference wage for married men of pensionable age. This amount was raised Introduction, Summary, and Policy Conclusions 57 to 45 percent in 1968. The European Code of Security of 1990 sets a mini- mum standard for members of the Council of Europe equal to 65 percent for married people of a specific age. 8. As a proxy for the poverty line, this study uses 35 percent of the average net wage, which very broadly approximates a US$2.25-a-day poverty line con- verted into national currency, adjusted for purchasing power parity, expressed relative to the national average net wage, and averaged across the eight study countries. 9. The APEX model was developed by Axia Economics, with funding from the OECD and the World Bank. The model codes detailed eligibility and benefit rules for first- and second-pillar schemes based on available public information that has been verified by country contacts. Because the details of the rules sometimes change on short notice (and limited public disclo- sure), the calculations presented here should be considered as best approx- imations only. 10. Gross replacement rates are simulated replacement rates for an unmarried male working a hypothetical career path under the assumption that real wage growth is 2 percent, inflation is 2.5 percent, the rate of return on invested assets is 3.5 percent, and the worker retires at the statutory retirement age. A 1.5 percentage point markup of net returns over the growth rate of average earnings is a conservative assumption in light of historical performance. The observed net rates of return over GDP growth (used as a proxy for earnings growth) in developed countries over the period 1970­1995 was 1.4 percent. For emerging economies (over a more recent period), the observed markup was 2.8 percent (Musalem and Bebczuk, 2009, forthcoming). 11. Only middle-income, partial-career workers are examined, because replace- ment rates are roughly comparable for workers with lower or higher levels of preretirement income. 12. For Romania, the calculations in figure 1.6 assume that first-pillar benefits are indexed using a combination of wages (with a 70 percent weighting) and prices (with a 30 percent weighting)--as a proxy for endogenous indexation, which depends on changes made to point values--while second pillar bene- fits are price indexed. 13. The projections for Poland's NDC scheme may refer only to old-age pensions rather than include disability pensions and survivorship benefits as well. As a result, the size and path of the deficit may be understated. 14. See Holzmann, MacKellar, and Repansek (2009) for a discussion of these issues for the countries of southeastern European. 15. This estimate is based on the assumption that real wage growth is 2 percent, the net real rate of return on invested assets is 3.5 percent, and benefits (both from the unfunded and funded pillars) are price indexed. 58 Adequacy of Retirement Income after Pension Reforms Bibliography Augusztinovics, M., and J. Köllő. 2009. "Decreased Employment and Pensions." In Pension Reform in Southeastern Europe: Linking to Labor and Financial Market Reforms, ed. R. Holzmann, L. MacKellar, and J. Repansek, pp. 89­104. Washington, DC: World Bank. Barr, N., and M. Rutkowski. 2004. "Pensions." In Labor Markets and Social Policy in Central and Eastern Europe: The Accession and Beyond, ed. Nick Barr, pp. 135­70. Washington, DC: World Bank. Council of Europe 1990. European Code of Social Security (Revised). Rome. European Commission. 2007. Pension Schemes and Projection Models in EU-25 Member Countries. European Economy Occasional Paper 37, Economic Policy Committee and Directorate General for Economic and Financial Affairs, Brussels. Gubbels, J., D. Snelbecker, and L. Zezulin. 2007. The Kosovo Pension Reform: Achievements and Lessons. Social Protection Discussion Paper 0707, World Bank, Washington, DC. Hinz, R., A. Zviniene, and A. Vilamovska. 2005. The New Pensions in Kazakhstan: Challenges in Making the Transition. Social Protection Discussion Paper 0537, World Bank, Washington, DC. Holzmann, R. 1997a. Pension Reform, Financial Market Development, and Economic Growth: Preliminary Evidence from Chile. International Monetary Fund Staff Paper 44 (June): 149­78. ------. 1997b. "Starting over in Pensions: The Challenges Facing Central and Eastern Europe." Journal of Public Policy 17 (3): 195­222. ------, ed. 2009. Aging Populations, Pension Funds, and Financial Markets: Regional Perspectives and Global Challenges for Central, Eastern, and Southern Europe. Washington, DC: World Bank and ERSTE Foundation. Holzmann, R., and R. Hinz. 2005. Old-Age Income Support in the 21st Century. Washington, DC: World Bank. Holzmann, R., L. MacKellar, and J. Repansek, eds. 2009. Pension Reform in South- eastern Europe: Linking to Labor and Financial Market Reforms. Washington, DC: World Bank. Holzmann, R., and E. Palmer, eds. 2006. Pension Reform: Issues and Prospects for Non-Financial Defined Contribution Schemes. World Bank: Washington, DC. ILO (International Labor Organization). 1952. ILO Convention 102. Geneva. ------. 1967. ILO Convention 128. Geneva. Musalem, A., and R. Bebczuk. 2009, Forthcoming. "Can the Financial Markets Generate Sustained Returns on a Large Scale?" In Aging Populations, Pension Funds, and Financial Markets: Regional Perspectives and Global Challenges for Introduction, Summary, and Policy Conclusions 59 Central, Eastern, and Southern Europe, ed. R. Holzmann. Washington, DC: World Bank and ERSTE Foundation. Nickel, J., and C. Almenberg. 2006. "Ageing, Pension Reforms and Capital Market Developments in Transition Countries." Working Paper, European Bank for Reconstruction and Development, London. OECD (Organisation for Economic Co-operation and Development). 2001. Ageing and Income: Financial Resources and Retirement in Nine OECD Countries. Paris: OECD. Orenstein, M. 2008. Privatization Pensions: The Transnational Campaign for Social Security Reform. Princeton, NJ: Princeton University Press. Rudolph, H., and R. Rocha. 2009, forthcoming. "Population Aging and the Payout of Benefits: Issues and Options." In Aging Populations, Pension Funds, and Financial Markets: Regional Perspectives and Global Challenges for Central, Eastern, and Southern Europe, ed. R. Holzmann. Washington, DC: World Bank and ERSTE Foundation. Schwarz, A. 2007. "Pensions." In Fiscal Policy and Economic Growth: Lessons for Eastern Europe and Central Asia, ed. Cheryl Gray, Tracey Lane, and Aristomene Varoudakis, Washington, DC: World Bank. Tesliuc, E., D. Coady, M. Grosh, and L. Pop. 2008. Program Implementation Matters for Targeting Performance: Evidence and Lessons from Eastern and Central Europe. Washington, DC: World Bank. UN (United Nations). 2007. World Population Prospects, 2006 Revision. New York: United Nations. Valdés-Prieto, S. 2008. Designs for the First-Pillar Pensions and the 2008 Chilean Reform. http://editorialexpress.com/cgi-bin/conference/download.cgi?db_ name=SECHI2008&paper_id=130. Whitehouse, E. 1999. Tax Treatment of Funded Pensions. Washington, DC: World Bank. World Bank. 1994. Averting the Old-Age Crisis. Washington, DC: World Bank. ------. 2005. Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union. Washington, DC: World Bank. ------. 2007a. From Red to Gray: The "Third Transition" of Aging Populations in Eastern Europe and the Former Soviet Union. Washington, DC: World Bank. ------. 2007b. Pensions Panorama. Washington, DC: World Bank. C H A P T E R 2 Bulgaria Bulgaria inherited a socialist-era defined-benefit pension system financed on a pay-as-you-go basis (meaning that contributions from current workers are used to pay benefits to current beneficiaries). The country's transition from central planning to a market economy affected all sectors of the economy. The change caused living standards for the elderly to decline and increased fiscal pressure on the pension system. In 1991, the third year of transition, the pension system generated a deficit equivalent to 2.96 percent of gross domestic product (GDP), largely as a result of declining revenues from higher levels of unemploy- ment and growing informality in the labor markets. Between 1991 and 1999, the pension system generated deficits of 0.7­3.1 percent of GDP (except in 1997, when it barely broke even), highlighting the vulnera- bility of the system to short-term economic changes. In response to these deficits--and to the fact that the system will face even greater fiscal pressure as a result of the aging of the population-- the government launched a comprehensive reform of the pension system in 2000. The reform program included both the redesign of the existing pay-as-you-go scheme and the introduction of a privately managed, fully funded defined-contribution scheme. 61 62 Adequacy of Retirement Income after Pension Reforms Despite these reforms, the pension system is projected to generate deficits reaching 2.3 percent of GDP by 2050. Improving the long-term finances of the pension system will require that benefits be made less generous, that retirement ages be raised, or both. This trade-off between the financial sustainability of the pension system and the benefits it pro- vides will become even more pronounced as life expectancy increases. Against this backdrop, this chapter evaluates Bulgaria's pension sys- tem, focusing on fiscal sustainability and benefit adequacy. Adequacy is evaluated through the lens of statutory net replacement rates for differ- ent retirement ages, patterns of contributions, and income levels relative to international benchmarks. This chapter is organized as follows. The next section discusses the motivation for the reforms. The following section describes the key char- acteristics of the reformed pension system. The third section assesses the adequacy of pension benefits and the fiscal sustainability of the system. The last section draws conclusions. Motivation for Reform Bulgaria inherited a socialist-era pension scheme that suffered from a number of serious design flaws, including relatively high contribution rates, an unequal distribution of the insurance burden, and early retire- ment (Shopov 1998, 2001). These flaws became increasingly evident in the 1990s, during the country's transition from central planning to a mar- ket economy. As a result of declining formal sector employment and increasing informality in the labor markets, pension system revenues fell, resulting in a pension system deficit of 2.96 percent of GDP in 1991 (table 2.1). Between 1991 and 1999, the pension system generated deficits of 0.7­3.1 percent of GDP (except for 1997, when it barely broke even), as a result of large and unpredictable fluctuations of both revenues and expenditures. Over this period, the number of contributors fell 28 percent, while the number of pensioners increased 5 percent. Projections indicate that, in the absence of reform, the pension scheme would have generated deficits equivalent to 3.4 percent of GDP by 2050 (figure 2.1). The aging of the population is apparent from projections of the country's old-age dependency ratio (the population age 65 and older divided by the population age 20­64), which is projected to increase from 27.5 percent in 2005 to 65.8 percent by 2050 (figure 2.2). In response to these deficits--and recognizing that the pension system will face greater fiscal pressure as a result of the aging of the Bulgarian Bulgaria 63 Table 2.1 Fiscal Balance of Bulgaria's Pension System before Reform, 1990­99 (percentage of GDP) Year Revenuesa Expenditures Balance 1990 10.98 10.92 0.06 1991 9.03 11.99 ­2.96 1992 11.03 12.31 ­1.28 1993 10.35 13.48 ­3.13 1994 8.90 11.47 ­2.57 1995 7.93 9.18 ­1.25 1996 7.10 7.81 ­0.71 1997 7.53 7.53 0.00 1998 8.66 9.80 ­1.14 1999 8.66 10.23 ­1.57 Source: National Social Security Institute 2005. a. Includes contribution revenues only. Figure 2.1 Projected Fiscal Balance of Bulgaria's Public Pension System before Reform, 2001­50 12 10 8 6 GDP of 4 2 0 percentage ­2 ­4 ­6 2001 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 year expenditures revenues deficit Source: National Social Security Institute 2005. population--the government launched a comprehensive reform of the system in 2000. This reform included the redesign of the existing pay-as- you-go scheme; the introduction of a privately managed, fully funded defined-contribution scheme; and the shifting of early retirement for workers engaged in hazardous occupations from first-pillar to second- pillar occupational pension schemes sponsored by employers (Hristoskov 64 Adequacy of Retirement Income after Pension Reforms Figure 2.2 Projected Old-Age Dependency Ratio in Bulgaria, 2005­50 70 60 (percent) 50 ratio 40 30 dependency 20 10 old-age 0 2005 2010 2020 2030 2040 2050 year Source: Reiterer 2008. 2000, 2002). This reform moved the pension system from a monopillar design based solely on pay-as-you-go financing to a multipillar design. The new system also includes a voluntary, fully funded defined-contribu- tion scheme, introduced in 1994. Characteristics of Bulgaria's Pension System This section describes the main characteristics of Bulgaria's pension sys- tem. These include the design of the individual pillars of social insurance; the rules governing pension system taxation, institutional structure, and coverage; and the provisions governing old-age, disability, and survivorship pensions. The design of the pension system is assessed using a conceptual framework developed by the World Bank, which generally recommends including a funded component if conditions are appropriate but increas- ingly recognizes that a range of choices is available to policy makers to pro- vide effective old-age protection in a manner that is fiscally responsible (see Holzmann and Hinz 2005). In general, the World Bank supports pension systems composed of some combination of five basic pillars: · a noncontributory (or zero) pillar (in the form of a demogrant, social pension, or social assistance benefit) intended to provide a minimal level of income protection; Bulgaria 65 · a first-pillar contributory system linked to earnings, which seeks to replace a portion of preretirement income; · a mandatory second pillar (essentially, individual savings accounts), which can be designed in various ways; · a voluntary third pillar, which is flexible and discretionary (this pillar, too, can take a variety of forms); and · a fourth pillar of informal intrafamily or intergenerational sources of financial and nonfinancial support to the elderly, including access to health care and housing. Pillar Design The design of Bulgaria's pension system incorporates all five of the pillars recommended by the World Bank (table 2.2). The publicly managed non- contributory zero pillar, financed with general tax revenues, redistributes income to lower-income groups using means testing, so that eligible ben- eficiaries receive a benefit sufficient to ensure them a total income equal to the state-defined minimum income guarantee. Both the traditional publicly managed pay-as-you-go first pillar and the privately managed, fully funded second pillar are earnings-related schemes. First-pillar benefits are computed taking into account the length of an individual's service and the individual's earnings throughout his or her working life. Benefits are adjusted in retirement using a formula based on the average of inflation and wage growth (Swiss indexation), thereby allowing pensions to rise more rapidly than inflation without imposing the heavier fiscal burden of wage indexation. Second-pillar benefits are determined by an individual's contributions and investment earnings. At retirement, account balances are converted to income streams by the pension fund rather than used to purchase annuities. Supplementing these earnings-related schemes is a voluntary privately managed third pillar, which is intended to provide individuals with a mechanism for supplementing the benefits provided by the mandatory pillars. The fourth pillar provides health care to the elderly as part of the overall health-care system. The mandatory first and second pillars are completely tax exempt, a policy that is uncommon. Most countries impose some taxation, at the point at which pension contributions are made, investment income is earned (for funded schemes), or benefits are received (see box 1.1 in chapter 1). The third pillar is subjected to an exempt-exempt-taxed regime, meaning that contributions are partially exempt from taxation, investment income is fully exempt, and benefits are taxed. Table 2.2 Structure of the Bulgarian Pension System 66 Taxation Investment income/ Benefit capital Scheme type Coverage Type Function Financing Generic benefit indexation Contributions gains Benefits Zero pillar (public Universal Means tested Redistribution Tax revenue Defined every Ad hoc decisions n.a. n.a. Exempt noncontributory) year by law by the government First pillar (public, Mandatory Defined Insurance Percentage of Benefit calculated 50 percent Exempt n.a. Exempt earnings related) benefit individual on the basis of inflation, earnings contribution 50 percent period and average individual insured wage coefficient growth over previous year Second pillar Mandatory Defined Insurance Percentage of Pension from Income stream Exempt Exempt Exempt (private, earnings contribution individual capital from related) earnings accumulation conversion of capital accumulation Third pillar Voluntary Defined Insurance Voluntary Pension from Depends Exempta Exempt Taxed (private, contribution contributions capital on options voluntary) accumulation chosen Fourth pillar Mandatory n.a. Insurance Percentage Specified basic n.a. Exempt n.a. n.a. (public health of individual health service care) earnings plus package tax revenues Sources: European Commission 2007; Shopov 2007. n.a. = Not applicable. a. Contributions up to 10 percent of earnings are exempt from taxation; monthly employer contributions of up to lev 60 per worker are exempt from the corporate income tax base. Bulgaria 67 Noncontributory scheme. Bulgaria provides a noncontributory social pen- sion to people over 70 who are not collecting a pension. Eligibility requires that average income per family member be lower than the guaranteed min- imum income for a 12-month period. The government determines the amount of the social pension annually. Since January 2006, the old-age social pension has been roughly 63 leva (Lev) (17.7 percent of the average wage). In 2005, 4,592 people received social pensions (roughly 0.20 per- cent of the total number of pensioners), a relatively low percentage that reflects the artificial full employment of the socialist era, which qualifies most elderly people for a pension from the first-pillar scheme. Another noncontributory source of income support comes from the guaranteed minimum income (GMI) program, which provides a means- tested benefit. An important part of the overall social assistance program, the GMI is accessible to the entire population, including the elderly. The amount of GMI was defined in 1992 using an established minimum con- sumption basket. Since then, it has been increased on the basis of budg- etary resources rather than any sort of indexation rule. In 2005, the GMI was Lev 55 (roughly 37 percent of the minimum wage). The benefits paid under the GMI are adjusted in value such that the beneficiary's total income attains the minimum threshold, which depends on household size, thereby awarding higher benefits to larger households. Retirement benefits are considered when determining the amount of the benefit. In 2005, total expenditures attributable to the GMI were equiva- lent to 0.26 percent of GDP; 21,600 elderly people (about 10 percent of all GMI beneficiaries) received benefits. Earnings-related schemes. Both the traditional publicly managed pay-as- you-go first pillar and the privately managed, fully funded second pillar are earnings-related schemes. The first pillar was reformed in 2000; the second pillar was introduced in 2002 (table 2.3). Retirement ages were gradually increased from 60 years to 63 years for men (to be implemented by 2005) and from 55 years to 60 years for women (to be implemented by 2009). Eligibility was made conditional on the accumulation of qualification points, defined as the sum of an individual's age and years of contribu- tions. The income on which benefits are computed was changed from the highest 3 of 15 years of earnings to lifetime earnings (with a grandfather clause for earnings for years before 1997), thereby strengthening the link between contributions and benefits. The 2000 reforms also introduced two types of fully funded second- pillar pension schemes: occupational schemes, which were launched in 68 Table 2.3 Parameters of Earnings-Related Schemes in Bulgaria before and after Reform Vesting Contribution Contribution Pension Retirement Scheme type Period period rate ceiling Benefit rate assessment base age First pillar Prereform 10 years 39 percent (37 percent n.a. 55 percent of the wage Highest 3 of 60 for men, (earnings by employer, 2 percent base (individual last 15 years 55 for women related, by employee) coefficient times the universal) average monthly wage for the 3 preceding years) Postreform 15 years 23 percenta (14.95 Annually fixed in 1 percent per year Highest 3 of last 63 for men, percent by employer, budget law 15 years before 1997; 60 for women 8.05 percent by (Lev 1,400 in 2007, entire working period employee) roughly 3.5 times after 1997 average wages) Second pillar Prereform n.a. n.a. n.a. n.a. n.a. n.a. (earnings Postreform n.a. 5 percent (3.25 percent Annually fixed Pension from capital Accumulated 63 for men, 60 for related, by employer, 1.75 in the budget accumulation funds women (58 for men universal) percent by employee) law (Lev 1,400 and 55 for women in in 2007) case of unemployment) Second pillar Postreform 10 years 7­12 percent, paid Annually fixed Timebound early Accumulated 55 for men and 52 for (earnings for Category entirely by employer in budget law retirement pension from funds women for Category I, related, I, 15 years for (Lev 1,400 capital accumulationb 60 for men and 57 for occupational) Category II in 2007) women for Category II Sources: European Commission 2007; Shopov 2007. n.a. = Not applicable. a. This figure represents the total contribution for old-age, disability, and survivor pensions in 2007. Individuals participating in only the first pillar pay 23 percent (14.95 percent employer, 8.05 employee). Individuals participating in both the first and the second pillars pay 18 percent (11.70 percent by employer, 6.30 percent by employee) to the first pillar and 5 percent (3.25 percent by employer, 1.75 percent by employee) to the second pillar. b. This accumulation finances a bridging pension for certain occupations between their low retirement age and the normal retirement age. Bulgaria 69 2000, and universal (or open) schemes, which were launched in 2002 and are mandatory for people born after December 31, 1959. Second-pillar benefits are a function of an individual's contributions and investment earn- ings. At retirement, account balances are converted into income streams by the pension fund rather than used to purchase annuities.1 Both universal earnings-related schemes are financed by contributions from employees and employers; occupational earnings-related schemes are financed by employers only. Before the 2000 reforms, the contribution rate was 39 percent--a comparatively high levy. Because this rate was believed to negatively affect labor competitiveness and create incentives for evasion, the 2000 reforms reduced the rate to reach 23 percent in 2007. Five percent- age points of this amount are diverted to the second-pillar universal pension scheme; the remainder is used to finance benefits under the first pillar. Voluntary scheme. Bulgaria introduced a voluntary privately managed third-pillar pension scheme in 1994 (table 2.4). The scheme is open to everyone 15 and older, with or without an established employer rela- tionship. Contributions must be at least 10 percent of the minimum wage. Actual contributions are determined by a contract with the pension insurance company; they can be set as an absolute amount or as a per- centage of the minimum wage or the participant's earnings. When the amount of contributions changes, the contract must be amended. Participants may contribute monthly or at other intervals, depending on the rules of the fund. Lump-sum contributions in larger amounts are also possible. Participation in the scheme is promoted using tax policy. Ten percent of total contributions are fully exempt from taxation. Employers who contribute on behalf of employees may deduct up to Lev 60 per employee per month from their taxable income. Participants are entitled to benefits upon reaching age 58 for men or 55 for women. Participants can collect benefits for five years before reaching retirement age if they are entitled to a benefit from the first-pillar scheme. Disability pensions are provided to people who become incapacitated. In 2006, eight pension fund management companies operated in the market, two of which managed 71 percent of total assets.2 In 2005, 549,851 people (10.3 percent of Bulgaria's working-age popu- lation) participated in the scheme. At the end of 2006, assets totaled Lev 497.9 million (1.0 percent of GDP). Health care system. Health care in Bulgaria is provided primarily through mandatory health insurance, although voluntary health insurance is available 70 Adequacy of Retirement Income after Pension Reforms Table 2.4 Characteristics of the Voluntary Scheme in Bulgaria Contributions Lump-sum Tax tax payments Vesting Retirement advantages to deductible by possible in Coverage period age participants employers retirement Open to No 58 for men, Yes Yes Yes anyone over 55 for age 16 women Source: Shopov 2007. to supplement the benefits of the mandatory system. The mandatory system covers roughly 92 percent of the population and is managed by the National Health Insurance Fund through contractual relationships with health care providers. The system is financed by contributions from economically active people and from the government on behalf of chil- dren under 18 and people receiving social assistance and pensions, among others. In 2007, the contribution rate was 6 percent, evenly split between employers and employees.3 Contributions paid on behalf of pensioners are based on the amount of their pensions. Pensioners are eli- gible for the same services as contributors to the health insurance sys- tem. The system is also financed by copayments equal to 1 percent of the minimum wage (Lev 180 in 2007) paid upon each visit to a general practitioner or outpatient specialist. Copayments for hospital stays are 2 percent of the minimum wage per day of care, with a ceiling of 10 days a year. Minors, unemployed family members, disabled military person- nel, medical staff, prisoners, and people eligible for social assistance (including the elderly receiving social assistance benefits) are not required to make copayments. In 2005, health expenditures accounted for 7.7 percent of GDP, 60.6 percent of which was public expenditures and 39.4 percent was private expenditure. Of private expenditure, 96.3 percent was attrib- utable to out-of-pocket expenditures (informal payments, direct pay- ments, and copayments) (WHO 2008). Institutional Structure and Coverage of Earnings-Related Schemes Bulgaria's mandatory pension schemes cover all salaried employees and self-employed persons, including farmers. The National Social Security Institute (NSSI) administers the first-pillar scheme and various non- contributory pensions through a central office and 28 regional offices. Contributions for the first and second pillars and for health insurance used Bulgaria 71 to be collected by the NSSI. Since January 2006, responsibility for col- lecting and transferring these contributions has been assumed by the National Revenue Agency, which collects most taxes in Bulgaria. In 2005, some 2.6 million participants (48.3 percent of the working- age population and 83.5 percent of the labor force) contributed to the first pillar. Of these participants, 2.2 million (84.6 percent of all first- pillar contributors) also participated in the mandatory second pillar and voluntary third pillar.4 Eight pension fund management companies operate in the mandatory pension fund market, which is overseen by the Financial Supervision Commission. At the end of 2006, assets totaled Lev 1,024.5 million (2.1 percent of GDP), of which 42 percent was managed by the two largest pension fund companies. Structure of Benefits The earnings-related pension scheme provides old-age, disability, and sur- vivorship pensions. The provisions governing each of these types of ben- efits are discussed as follows.5 Old-age benefits. Eligibility for an old-age pension in Bulgaria is condi- tional on the accumulation of qualification points, defined as the sum of an individual's age and years of contributions. Men must accumulate 100 points (37 years of service for men retiring at the normal retirement age), while women must accumulate 94 points (34 years of service for women retiring at the normal retirement age once the age reaches 60 in 2009). The pension assessment base (the wages used in computing benefits) is the product of an individual's coefficient and the average covered earn- ings over the preceding 12 months.6 Benefits accrue at the rate of 1 percent of the assessment base per year of contributory service. Starting in January 2007, the accrual rate for each year of service beyond 37 for men and 34 for women was 1.5 percent, an amount that increased to 3 percent in January 2008. Computing an individual's pension requires multiplying the pension assessment base by the individual's total accrual, which is a function of the individual's length of service. This formula can be presented as follows: old-age pension = IC * AMII (12 * CP * 1 percent) where IC = the individual coefficient of the pensioner, AMII = the average monthly insurable income over the 12 months before the individual was awarded a pension,7 and CP = the contribution period (in years). 72 Adequacy of Retirement Income after Pension Reforms Early retirement is not permitted, except for certain categories of workers, such as the military and police, who may retire at an earlier age with fewer years of service. Late retirement is available to everyone with- out restriction. Subject to eligibility criteria, pensioners may be eligible for benefits under the guaranteed minimum old-age pension, the amount of which is defined in the social security budget each year.8 People who do not meet these eligibility requirements are eligible for a minimum pension that is 85 percent of the minimum old-age pension, provided they have at least 15 years of credited contributory service and have reached age 65. In 2006, the amount of this pension was Lev 85 (24 percent of the average wage). The maximum pension provided by the first pillar cannot exceed 35 percent of the maximum insurance income for the preceding year. Second-pillar benefits are a function of an individual's contributions, invest- ment earnings, and life expectancy at retirement. Account balances are converted into income streams by the pension fund rather than used to purchase annuities.9 Disability benefits. Disability benefits are available to people who elected to remain only in the first-pillar scheme as well as to people who elected to participate in the new two-pillar scheme (table 2.5). Eligibility depends on service, with longer service requirements for older workers. Disability ben- efits are computed in a manner similar to that used to compute old-age pensions. Service credit is awarded for years lost to disability up to the nor- mal retirement age, with a coefficient applied to these years that reflects the degree of impairment. For people with more than 90 percent impairment, the coefficient is 0.9; for 71­90 percent impairment, the coefficient is 0.7; for 50­70 percent impairment, the coefficient is 0.5. The minimum disability benefit Table 2.5 Eligibility Conditions for and Benefits Provided by Disability Pensions under the First Pillar Earnings-Related Scheme in Bulgaria Vesting period Contribution rate Eligibility Benefit rate Partial pension Under age 20: No No specific At least 50 1 percent per Based on minimum contribution percent loss year degree of required rate of working disability Age 20­25: 1 year capacity Age 25­30: 3 years Over age 30: 5 years Sources: European Commission 2007; Shopov 2007. Bulgaria 73 ranges from 85 to 115 percent of the minimum old-age pension, depending on impairment. Under the second pillar, benefits are paid for life on the basis of the accumulated capital in an individual's account and life expectancy. Benefits are paid directly by the pension fund; annuities are not purchased. Survivor benefits. Survivor benefits are awarded to the dependents of individuals who, at the time of their death, were receiving (or had met the criteria to receive) an old-age or disability pension (table 2.6). Eligible survivors include widows and widowers who are unable to work or are within five years of the retirement age, orphans up to the age of 18 (26 if attending school), and parents who had been supported by the deceased. The benefit depends on the number of survivors in the deceased's household. It is 50 percent of the deceased's benefit for one survivor, 75 percent for two survivors, and 100 percent for three or more survivors, divided equally among all survivors. The minimum benefit is 75 percent of the minimum old-age pension (Lev 85 in 2006). Under the second pillar, survivors of working individuals receive the deceased's account balance paid as a lifetime annuity based on the accu- mulated capital in the deceased's account and life expectancy. The allo- cation among survivors is governed by the inheritance law. Benefits are paid directly by the pension fund; annuities are not purchased. Assessment of the Performance of Bulgaria's Pension System The World Bank has established four principles for evaluating public pen- sion systems, which together should guide the process of pension reform (see Holzmann and Hinz 2005). Broadly speaking, these include the adequacy and security of benefits, the affordability of contributions, the sustainability of the system over time, and the robustness of the system in the face of demographic changes and macroeconomic shocks. This section focuses primarily on the adequacy of benefits and the financial sustain- ability of the first- and second-pillar earnings-related pension schemes. The remaining principles are mentioned only briefly. Adequacy is analyzed through the lens of net replacement rates. Financial sustainability is evalu- ated using projections of pension expenditure and revenues. Benefit Adequacy Replacement rates are a useful yardstick for measuring the adequacy of pension benefits, because they express benefits relative to preretirement 74 Table 2.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in Bulgaria under the First-Pillar Earnings-Related Scheme Orphan Spouse replacement Benefit Remarriage Orphan replacement Total family Eligibility rate duration test age limit rate benefit Eligibility of the 1 survivor: 50 percent For life, if spouse Pension ceases 18 (26 if orphan 1 survivor: 100 percent deceased of deceased's benefit is disabled or if survivor is a student) 50 percent of regardless of for old-age 2 survivors: 75 percent has reached remarries deceased's number of or disability of deceased's benefit retirement benefit survivors; if pension 3 or more survivors: age 2 survivors: pension is less 100 percent of 75 percent of than minimum deceased's benefit deceased's pension, Alternatively, survivors benefit minimum can receive 20 percent 3 or more survivors: pension is paid of the pension of the 100 percent of deceased as a deceased's benefit supplement to their own pensions. Sources: European Commission 2007; Shopov 2007. Bulgaria 75 earnings, thereby indicating the degree to which income is replaced when workers retire. Two variants are commonly used. Gross replacement rates compute income replacement as the ratio of benefits paid to pretax pre- retirement earnings. Net replacement rates compute income replacement as the ratio of benefits received (that is, after the payment of taxes and other levies, including contributions for social insurance) to posttax pre- retirement earnings. In general, net replacement rates are a more useful measure of benefit adequacy, because they capture the degree to which actual take-home pay is replaced when workers retire. The level of income replacement at retirement is not the only measure of benefit adequacy. To fully assess benefit adequacy, it is also important to determine how postretirement indexation rules will affect replacement rates during retirement. Pension benefits in retirement are expected to be indexed to inflation, so that their real value is maintained. In a growing economy with rising real wages, however, mere price indexation of pen- sions leads to a deterioration of the relative consumption position of the retirees. For this reason, some countries have introduced mixed indexation of pensions that use varying weights of inflation and wage growth in the indexation formula. In order to evaluate the effect of indexation on replacement rates in Bulgaria, the replacement rates are normalized to 100 percent, and the assumptions for calculating the replacement rates are maintained (that is, inflation is 2.5 percent a year and real wage growth is 2.0 percent a year). The change in the replacement rate is measured in comparison to full wage indexation or the earnings of an active worker. The results of this analysis indicate that the relative income position of a retiree would deteriorate by 16 percent after 10 years in retirement and by 45 percent after 35 years in retirement. (The evaluation of income replacement that follows considers replacement rates only at retirement; it does not take into account the impact of indexation policies on replacement rates during retirement.) Replacement rates are a function of the formula governing pension benefits; an individual's contribution history; and, in the case of net replacement rates, the rules of income tax, social security contributions, and other relevant levies. The benefit formula establishes the degree to which the system redistributes income across individuals of different levels of preretirement earnings. Progressive systems provide higher lev- els of income replacement to people with lower levels of preretirement income. In general, the degree to which a system is redistributive depends on the existence (and value) of flat transfers and minimum pension 76 Adequacy of Retirement Income after Pension Reforms guarantees, the degree to which benefits are earnings related, and the existence of ceilings on earnings subject to contributions. An individual's contribution history can be characterized by his or her age of entry into the labor force, contribution density, and decisions regarding the timing of retirement. To some degree, these three factors are influenced by the incentives embodied in the pension system. The tax and contribution system influences net replacement rates through the progressiveness of the income tax formula, which taxes (higher) income during a worker's active life more than it taxes (lower) pension benefits in retirement. In addition, social security levies (for pensions; unemployment; health care; and, at times, housing and family benefits) are typically reduced or eliminated altogether in retirement. These benefits are particularly important for low- to middle-income groups. Benchmarks need to be established for the evaluation of the adequacy of the income replacement provided by the earnings-related pension schemes. Unfortunately, there is no consensus on what constitutes ade- quacy. According to one widely respected definition, pensions are ade- quate when they are sufficient to prevent poverty among the elderly and provide the vast majority of the population with a reliable mechanism for smoothing income over a lifetime. Even with a definition, however, establishing benchmarks is problematic, because attitudes vary across countries as a result of social and cultural perceptions. Moreover, bench- marks ignore the other factors that affect the welfare of the elderly--and that also vary across countries--including the existence and generosity of health insurance and long-term care, the cost of housing, the structure of traditional living arrangements, the presence of informal intrafamily or intergenerational sources of financial and nonfinancial support, and the availability and security of other mechanisms for people to save for their own retirement. One reputable nine-country study (OECD 2001) observes that living standards are roughly comparable for people 10 years older than the normal retirement age and people 15 years younger than the normal retirement age when retirees have disposable income equal to roughly 80 percent of the disposable income of working-age people. In part, this is attributable to the fact that retirees have no work-related expenses (they do not have to commute or buy special clothing or uniforms, for exam- ple). This finding, however, does not imply that mandatory first-pillar pension schemes should actually target an 80 percent net replacement rate. To the contrary, in middle- and high-income countries, one can reasonably expect individuals to save for their own retirement--and the Bulgaria 77 empirical evidence suggests that, in practice, they do so.10 There is also some evidence to suggest that the ratio between pre- and postretirement income is somewhat independent of the income replacement mandate of the public pension system. Put simply, individuals tend to save more in countries with more modest mandates (and vice versa). Because Bulgaria has access to relatively well-developed financial mar- kets, it would seem reasonable to expect middle- and higher-income workers to save enough to finance at least 25 percent, if not closer to 50 percent, of this 80 percent income replacement target. Given this, three benchmarks are provided: a 40 percent net replacement rate (which implies that individuals would be expected to save enough to finance half of the total income replacement target); a 60 percent net replacement rate (which implies that individuals would be expected to finance a quar- ter of the target); and an 80 percent net replacement rate (which implies that individuals, most of whom would be low-income earners, would not be expected to contribute anything toward the target).11 In the following analysis, these benchmarks are used to evaluate the adequacy of benefits in Bulgaria compared with the average net replacement rate observed in 53 countries around the world, the average net replacement rate observed in selected countries in Europe and Central Asia, and the poverty line in Bulgaria.12 To estimate gross and net replacement rates, we use the Analysis of Pension Entitlements across Countries (APEX) model to consider two critical dimensions: earnings levels and contribution periods.13 This model generates estimates for replacement rates under steady-state assumptions (that is, as if the rules of the reformed pension scheme had been in place over the entire active life of the individual). Because life expectancies at retirement are projected to increase over time--which will affect the benefits paid by defined-contribution pension schemes-- a reference year must be chosen. For the purpose of this study, 2040 is used, because it provides a sufficiently long contribution period over which to approximate steady-state conditions. The first critical task is to investigate levels of income replacement across a relevant spectrum of income. Income is represented as a per- centage (50­200 percent) of average earnings. The second task is to investigate the impact on income replacement of differences in the duration, timing, and density of an individual's contribution history (where density refers to the percentage of time an individual actually contributes over a given period). To facilitate the presentation of these multidimensional results, replacement rates are computed as a function 78 Adequacy of Retirement Income after Pension Reforms of the age at which an individual exits the labor market. They are presented separately for full-career and partial-career workers. Replacement rates for full-career workers. Projected replacement rates for full-career workers are examined first. For the purpose of this analysis, a full career is defined as continuous employment from age 20 to the cur- rent normal retirement age of 63 for men. Gross replacement rates clearly show why the earnings-related pension schemes have been described as providing a strong link between benefits and contributions (figure 2.3). Irrespective of income, gross replacement rates are 74.8 percent (55.0 percentage points provided by the first pillar, 19.8 percentage points provided by the second pillar). The situation does not change significantly when taxes are taken into consideration (figure 2.4). High-income earners receive net replacement rates that are identical to low- and middle income earners. Examination of replacement rates for full-career workers indicates that these pensions are adequate (figure 2.5), with replacement rates for all levels of preretirement income higher than the high benchmark.14 This suggests that the pension system is effectively smoothing consumption from work into retirement for all full-career workers and that the objec- tive of poverty alleviation is being met. Levels of income replacement for Figure 2.3 Sources of Gross Replacement Rates in Bulgaria, by Income Level 80 70 60 (percent) 50 rate 40 30 replacement 20 10 gross 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (points) Source: APEX model. Note: Figure shows projected replacement rate for full-career worker in 2040 as approximation of steady-state conditions. Bulgaria 79 Figure 2.4 Sources of Net Replacement Rates in Bulgaria, by Income Level 100 90 80 (percent) 70 rate 60 50 40 replacement 30 of 20 10 share 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (points) taxes Source: APEX model. Note: Figure shows projected replacement rate for full-career worker in 2040 as approximation of steady-state conditions. almost all full-career workers in Bulgaria are higher than regional and world averages. The Bulgarian pension system provides relatively little redistribution from comparatively well-off individuals to those with lower levels of preretirement income. Replacement rates for partial-career workers. Not everyone works from age 20 to the statutory retirement age. Many individuals enter and exit the labor force (often at different ages and for different periods of time) and earn different wages while working (figure 2.6). To examine the ade- quacy of benefits for partial-career workers, we consider three stylized cases. These cases include career type A (defined as someone entering the labor force at age 25 who works continuously for a period of years before leaving the workforce at some point between the ages of 50 and 70 and then claims a benefit); career type B (identical to career type A, except that the worker enters the workforce at age 30 and leaves no earlier than age 55); and career type C (identical to career type A, except that the individual contributes in only three years out of four while in the labor force). In cases where the withdrawal from the formal labor market occurs before the statutory retirement age, the pension is claimed (and the replacement rate calculated) only at the later age. For withdrawals after the statutory retirement age, the ages coincide. 80 Adequacy of Retirement Income after Pension Reforms Figure 2.5 Net Replacement Rates for Male Full-Career Workers in Bulgaria, Europe and Central Asia, and the World, by Income Level 100 80 high benchmark (percent) 60 middle benchmark rate 40 low benchmark replacement net 20 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage Bulgaria average Europe and Central Asia average world average poverty line (percentage of average income) Source: Authors' calculations based on World Bank 2007 and the APEX model. Note: Figure shows projected replacement rate for full-career worker in 2040 as approximation of steady-state conditions. Several conclusions can be drawn from figure 2.6. First, leaving the workforce early can be costly. Someone retiring before reaching the retirement age may not receive levels of income replacement higher than even the lowest of the three benchmarks--and those leaving very early may receive levels of income replacement below the poverty line. Second, entering the workforce later in life is costly. Someone entering the workforce at the age of 30 receives a net replacement rate that is 3­21 percentage points lower than someone entering the workforce at age 25. Third, working intermittently is costly. Someone entering the workforce at the same age but who contributes only three years out of four will receive a net replacement rate that is 3­41 percentage points lower than someone who contributes continuously. In all cases, net replacement rates grow faster the longer someone continues to work. This is encouraging, because it indicates that the pension system provides incentives for peo- ple to remain in the workforce. Although career type A workers can attain Bulgaria 81 Figure 2.6 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Bulgaria, by Career Type and Exit Age high benchmark 100 world and Europe and Central Asia average 90 80 70 (percent) 60 rate middle bechmark 50 low benchmark 40 poverty line (percentage 30 of average income) replacement 20 net 10 0 45 50 55 60 65 70 exit age from labor market career type A: career type B: career type C: entry age--25, entry age--30, entry age--25, contribution contribution contribution density--100% density--100% density--75% Source: Authors' calculations based on World Bank 2007 and the APEX model. Note: Figure shows projected replacement rate for partial-career worker in 2040 as approximation of steady-state conditions. (See text for descriptions of career types.) the 60 percent benchmark before reaching the normal retirement age, career type B and C workers must work three and seven years, respectively, beyond the normal retirement age to attain this benchmark. To replace 80 percent of preretirement earnings, career type A workers must work three years past the normal retirement age. Career type B workers must work until age 70 to attain the 80 percent benchmark. Career type C workers cannot attain the 80 percent benchmark even if they work until age 70. Fiscal Sustainability The sustainability of a pay-as-you-go first-pillar pension scheme is best eval- uated in actuarial terms by estimating the scheme's actuarial deficit as the difference between its assets and liabilities. If a large actuarial deficit exists, the scheme is financially unsustainable and needs policy actions that increase its assets, reduce its liabilities, or both. A good proxy for the actu- arial deficit is the difference between the present value of the scheme's expected future revenues (that is, contributions and other income) and 82 Adequacy of Retirement Income after Pension Reforms expected future expenditures (that is, benefit payments, administrative costs, and other expenses) over an extended projection period. The differ- ence between these two values represents an unfunded liability (some- times referred to as a financing gap) on the public sector balance sheet. Because this study is also concerned with the time path of revenues and expenditures (and the resulting balance across the projection period end- ing in 2050), this more pragmatic approach has been taken. Projections of expenditures, revenues, and deficits are presented on the basis of available postreform fiscal projections. Despite the improvements attributable to the reforms of 2000, the Bulgarian pension system is expected to generate deficits into the foresee- able future (figure 2.7). Revenues are projected to stabilize at about 8­9 percent of GDP over the period 2001­50, reaching 8.4 percent of GDP by 2050. Expenditures are projected to increase from 9.7 percent of GDP in 2001 to 10.7 percent of GDP by 2050, as first-pillar benefits decline as a share of the total benefits provided by the first two pillars. The net result is a projected deficit of 2.3 percent of GDP in 2050. What options exist for restoring the system to fiscal balance? Unfortunately, for policy makers, the options are limited. Revenues can be increased by increasing the contribution rate. Alternatively--or in addi- tion, as the options are not mutually exclusive--expenditures can be reduced by cutting benefits, increasing the minimum number of years Figure 2.7 Projected Fiscal Balance of Bulgaria's Public Pension System after Reform, 2001­50 12 10 8 GDP 6 of 4 2 percentage 0 ­2 ­4 2001 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 year expenditures revenues deficit Source: National Social Security Institute 2005. Bulgaria 83 required to become eligible for benefits, or delaying the payment of benefits by raising the retirement age further. Because raising the con- tribution rate could threaten competitiveness and will likely strengthen incentives for tax evasion, it is typically not embraced (it would also represent a reversal of policy, because Bulgaria deliberately reduced the contribution rate since 2000 to dampen the adverse impact of high taxes on labor markets). This leaves cutting benefits, tightening eligibility con- ditions, or raising the retirement age. As a major part of the deficit reflects the transition deficit toward the second pillar, the government may also consider financing part or all of the transition deficit through general revenues. If it does otherwise, restoring sustainability may reduce the ade- quacy of benefits provided to future beneficiaries. (A back-of-the-envelope analysis suggests that by 2050, retirement ages would have to be increased to at least 69 for men and women in order to bring the system to long- term fiscal balance.15) If retirement ages are left unchanged and the current structure of the system is retained, further cuts in benefits--on the order of a 32 percent reduction in the average benefit provided under the first pillar--will be required to make the system sustainable. If benefits are adjusted to main- tain a similar fiscal balance in proportion to the overall size of the first- pillar scheme, full-career workers will receive replacement rates roughly 20 percentage points lower in 2040 than they receive today (figure 2.8). Two conclusions can be drawn from comparison of these new (and lower) net replacement rates against the three benchmarks. First, a 32 percent reduction in benefits results in income replacement broadly at the poverty line for low-income full-career workers. For middle- and high-income full-career workers, replacement rates are significantly higher than the poverty line. This indicates that the pension meets its poverty alleviation objective. Second, the same reduction in benefits will still support the objective of smoothing consumption for middle- and high-income full-career workers, because levels of income replacement are still higher than their 60 percent benchmark. This last observation is subject to three caveats. First, this analysis con- siders only full-career workers, while the average worker now contributes for only 27­30 years, substantially less than the 43 years expected of a full career. Contributing to the pension scheme for only 33 years, for exam- ple, reduces net income replacement by 30 percentage points for the average worker. Second, if benefit cuts are combined with further increases in the retirement age, benefit cuts will not need to be as steep in order to restore fiscal balance. Third, workers always have the option 84 Adequacy of Retirement Income after Pension Reforms Figure 2.8 Net Replacement Rates for Men in Bulgaria before and after Benefit Adjustment 90 80 high benchmark 70 (percent) 60 middle benchmark rate 50 40 low benchmark 30 replacement net 20 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage net replacement rate adjusted net replacement rate poverty line (percentage of average income) Source: Authors' calculations based on World Bank 2007 and the APEX model. Note: Figure shows projected replacement rate for full-career worker in 2040 as approximation of steady-state conditions. of saving outside of the first-pillar pension scheme. To increase income replacement by 1 percentage point, for example, a full-career worker would need to save only about 0.47 percent of his or her earnings from age 40 to the current age of retirement.16 Conclusions In response to deficits in the public pension system--and to the fact that the system will face even greater fiscal pressure as a result of the aging of the population--Bulgaria launched a comprehensive reform of the pen- sion system between 2000 and 2002. The reform program included both the redesign of the existing pay-as-you-go scheme and the introduction of a privately managed, fully funded defined-contribution scheme, which pro- vides workers with a mechanism for diversifying their retirement savings. Bulgaria 85 Together, these reforms strengthened the link between contributions and benefits and improved incentives for workers to remain in the workforce after reaching the minimum retirement age. The resulting gross and net replacement rates for full-career workers are well above the high bench- mark of 80 percent (and higher than regional and international bench- marks [see chapter 1]). As in other countries, workers with less than full careers--because they left the workforce before reaching retirement age, worked intermittently, or have gaps in their employment history--risk receiving income replacement that is closer to--or even below--the lower benchmark of 40 percent. As a result of reform, the long-term fiscal position of Bulgaria's pen- sion system has improved somewhat, with projected deficits falling from 3.9 percent of GDP in 2008 to 2.3 percent of GDP in 2050. Unless the government is willing and able to finance part or all of the transition deficit resulting from the introduction of the second pillar from general revenues, however, further improvements in the fiscal balance of the first- pillar scheme will require it to increase retirement ages further, to reduce benefits, or both. Increasing the retirement age in step with increases in life expectancy at retirement is a natural choice for both individuals and policy makers, but it requires cross-sectoral policy reforms to enable eld- erly workers to continue to participate in the labor market.17 If retirement ages are left unchanged, average initial replacement rates will have to be reduced by some 20 percentage points to restore the scheme to fiscal balance. Roughly half of these savings can be achieved by moving from Swiss indexation to price indexation of benefits. Following a reduction in benefits, the pension system would still meet its poverty alleviation objective, but it would not be able to adequately smooth income for low-income partial-career workers. Of course, indi- viduals have the option of participating in the voluntary fully funded third-pillar pension scheme, which was introduced to enable workers to save more for retirement than is provided by the mandatory schemes. This option is less relevant for low-income individuals, however, who have lower saving capacity. Notes 1. There appears to be sufficient political will to enact amendments governing the conversion of accumulated capital from the funded scheme into lifetime annuities, but the mechanism by which those annuities will be provided is still being debated. 86 Adequacy of Retirement Income after Pension Reforms 2. See the Web site of the Financial Supervision Commission (http://www.fsc.bg). Each company manages three separate pension funds (two that are mandatory and one that is voluntary). 3. The government had planned to increase the contribution rate for health care to 8 percent in 2008, but the increase was postponed until 2009. The ratio of the shares paid by employers and employees has changed and is expected to change further over time. The ratio was 80:20 in 2000, 75:25 in 2002­04, and 70:30 in 2005. Going forward, it will be 60:40 in 2008, 55:45 in 2009, and 50:50 from 2010 onward. 4. Some 66.5 percent of all workers participated in universal (open) funds, 3.8 percent participated in occupational (closed) funds, and 14.3 percent participated in voluntary pension funds in 2005. 5. These rules apply to new entrants; people with accrued rights are subject to transition provisions not addressed in this discussion. 6. An individual's coefficient is determined using the ratio of the individual's average contributory income over three consecutive years (chosen by the individual from the period 1982­96) to the national average salary over the same three-year period) and the ratio of the individual's average contributory income from 1997 onward to the national average monthly salary over the same period (European Commission 2007). 7. The NSSI calculates and publishes this amount every month. 8. The minimum pension is also the basis for the minimum disability pension and the minimum survivor pension. 9. The first old-age beneficiaries are not expected to begin drawing benefits until about 2018. Currently, the rules provide for phased withdrawals, because annuities have not yet been legislated (and only disability pensioners are currently receiving benefits). 10. In Chile, for instance, 70 percent of retirees from the mandatory public pension system own their home, which is a form of savings (see Valdés- Prieto 2008). 11. These benchmarks approximate the standards developed by the International Labour Organization (ILO) and the Council of Europe (1990). ILO Convention 102 of 1952 sets a minimum benefit equal to 40 percent of the reference wage for married men of pensionable age. This amount was raised to 45 percent in 1968. The European Code of Security of 1990 sets a minimum standard for members of the Council of Europe equal to 65 percent for married people of a specific age. 12. As a proxy for the poverty line, this study uses 35 percent of the average net wage, which very broadly approximates a US$2.25-a-day poverty line con- verted into national currency, adjusted for purchasing power parity, expressed relative to the national average net wage, and averaged across the Bulgaria 87 nine study countries. Such an approach enables valid comparisons to be made across the sample (see chapter 1). 13. The APEX model was developed by Axia Economics, with funding from the OECD and the World Bank. The model codes detailed eligibility and benefit rules for first- and second-pillar schemes based on available pub- lic information that has been verified by country contacts. Because the details of the rules sometimes change on short notice (and limited public disclosure), the calculations presented here should be considered as best approximations only. 14. Replacement rates are simulated for an unmarried male working a hypo- thetical career path under the assumption that real wage growth is 2 per- cent, inflation is 2.5 percent, the rate of return on invested assets is 3.5 percent, and the worker retires at the statutory retirement age. Replacement rates shown do not consider the benefits received from occupational schemes. 15. This estimate is based on the World Bank's baseline demographic projec- tions, which assume that everyone over the age of 68 receives a pension, everyone age 20­68 contributes, and all pensioners receive the replacement rate awarded from the first pillar to the median worker. 16. This estimate is based on the assumption that real wage growth is 2 percent, the net real rate of return on invested assets is 3.5 percent, and benefits (both from the unfunded and funded pillars) are price indexed. Country-specific mortality rates are used in this analysis. 17. See Holzmann, MacKellar, and Repansek (2009) for a conference volume that addresses theses issues for the countries of southeastern Europe. Bibliography Council of Europe. 1990. European Code of Social Security (Revised). Rome. Georgieva, L., P. Salchev, R. Dimitrova, A. Dimova, and O. Avdeeva. 2007. Health Systems in Transition: Bulgaria. European Observatory on Health Systems and Policies, Copenhagen. European Commission. 2007. Mutual Information System and Social Protection (MISSOC) database. http://ec.europa.eu/employment_social/. Financial Supervision Commission. 2005. Annual Report. Sofia. ------. 2006. Annual Activity Report. Sofia. GVG (Gesellschaft für Versicherungswissenschaft und -gestaltung) and European Commission. 2003. Study on the Social Protection Systems in the 13 Applicant Countries: Bulgaria Country Report. Cologne. Holzmann, R., and R. Hinz. 2005. Old Age Income Support in the 21st Century. Washington, DC: World Bank. 88 Adequacy of Retirement Income after Pension Reforms Holzmann, R., L. MacKellar, and J. Repansek, eds. 2009. Pension Reform in South- eastern Europe: Linking to Labor and Financial Market Reforms. Washington, DC: World Bank. Hristoskov, J. 2000. "Mandatory Social Insurance: Changes, Nature, and Content." In The Bulgarian Pension Model, USAID Bulgarian Pension Reform Project, Sofia. ------. 2002 "The Relations between the Public and Private Pension Insurance Systems: The Bulgarian and Foreign experience." Private Pension Series 4, Organisation for Economic Co-operation and Development, Paris. ILO (International Labour Organisation). 1952. ILO Convention 102. Geneva: ILO. ------. 1967. ILO Convention 128. Geneva: ILO. National Social Security Institute. 2005. Statistical Yearbook. Sophia: National Social Security Institute. OECD (Organisation for Economic Co-operation and Development). 2001. Ageing and Income: Financial Resources and Retirement in 9 OECD Countries. Paris: OECD. Reiterer, A. 2008. Population Development and Age Structure in Southeastern Europe until 2050. Social Protection and Labor Unit, World Bank, Washington, DC. Shopov, G. 1998. "Bulgarie: La reforme du systčme de retraite." Chronique Internationale de l'IRES 55. November. ------. 2001. "Bulgarian Pension System in Restructuring." In Ten Years of Economic Transformation. Vol. III. Studies in Industrial Engineering and Management 16. Lappeenranta University of Technology, Finland. ------. 2007. "The Bulgarian Pension System." Background paper prepared for this chapter. Institute of Economics, Bulgarian Academy of Sciences, Sofia. U.S. Social Security Administration. 2006. Social Security Systems throughout the World: Europe. Washington, DC: Social Security Administration. Valdés-Prieto, S. 2008. Designs for the First-Pillar Pensions and the 2008 Chilean Reform. http://editorialexpress.com/cgi-bin/conference/download.cgi?db_ name=SECHI2008&paper_id=130. World Bank. 1999. Pension Reform Strategy of Bulgaria. http://www.worldbank. org/pensions. ------. 2007. Pensions Panorama. Washington, DC: World Bank. WHO (World Health Organization). 2008. Database. http://www.who.int/ research/en/. C H A P T E R 3 Croatia Croatia inherited from the former Yugoslavia a traditional defined-benefit pension system financed on a pay-as-you-go basis (meaning that contri- butions from current workers are used to pay benefits to current bene- ficiaries). Within a few years of the country's transition to a market economy, pension expenditures began to increase, from a level equiva- lent to 9.7 percent of gross domestic product (GDP) in 1994 to 12.8 percent of GDP by 2000. Over this period, pension revenues declined from 8.1 percent of GDP to 7.3 percent of GDP, resulting in a pension deficit of 5.5 percent of GDP in 2000. Recognizing that these deficits were not sustainable and that the sys- tem would face even greater challenges in the medium to long term as the population ages, the government began a process of pension reform in 1995 that eventually replaced the traditional scheme with a three-pillar pension system--an approach to reform that was relatively common among transition economies. In 1998, the government replaced the tradi- tional defined-benefit formula with a new formula based on points. In 2002, it introduced a mandatory funded defined-contribution pension scheme and a voluntary scheme to supplement the two mandated schemes. As a result of these reforms, revenues are now projected to remain stable and expenditures to drop gradually, so that the deficit will become pro- gressively smaller by 2040. 89 90 Adequacy of Retirement Income after Pension Reforms Against this backdrop, this chapter evaluates the Croatian pension sys- tem, focusing on fiscal sustainability and benefit adequacy. Adequacy is evaluated through the lens of statutory net replacement rates for differ- ent retirement ages, patterns of contributions, and income levels, with comparisons to international benchmarks. The chapter is organized as follows. The next section discusses the motivation for the reforms. The following section describes the key char- acteristics of the reformed pension system. The third section assesses the adequacy of pension benefits and the fiscal sustainability of the system. The last section draws conclusions. Motivation for Reform Upon gaining independence in 1991, Croatia inherited a socialist-era public pension system financed on a pay-as-you-go basis. The system suf- fered from a number of serious design flaws similar to those observed in other transition economies, including low retirement ages, special privi- leges for certain occupations, and a high incidence of disability among war veterans. The fiscal balance of the pension system deteriorated steadily throughout the 1990s, with the deficit increasing from 1.6 percent of GDP in 1994 to 5.5 percent of GDP by 2000.The rising deficit was largely the result of steadily rising expenditures (table 3.1), although revenues did fall slightly between 1996 and 2000 as a result of reduced formal sector employment (caused partly by rising unemployment and partly by increas- ing labor market informality) and enterprise restructuring. The pension system was expected to face even greater challenges in the medium to long term, as the population ages. Croatia's old-age dependency ratio (the pop- ulation age 65 and older divided by the population age 20­64) is expected Table 3.1 Projected Fiscal Balance of Croatia's Public Pension System before Re- form, 1994­2000 (percentage of GDP) Year Revenues Expenditures Balance 1994 8.1 9.7 ­1.6 1995 8.6 10.8 ­2.2 1996 8.7 11.4 ­2.7 1997 8.6 12.5 ­3.9 1998 7.5 12.0 ­4.5 1999 7.4 13.3 ­5.9 2000 7.3 12.8 ­5.5 Source: Anusic, O'Keefe, and Madzarevic-Sujster 2003. Croatia 91 Figure 3.1 Projected Old-Age Dependency Ratio in Croatia, 2006­50 60 50 40 30 percent 20 10 0 2006 2016 2026 2036 2046 2050 year Source: Reiterer 2008. to increase substantially in the coming decades, from 27.8 percent in 2006 to 51.1 percent by 2050 (figure 3.1).1 Recognizing these challenges, the government began a process of pen- sion reform in 1995. In 1998, it eliminated the traditional defined-benefit formula used to calculate pensions in favor of a new formula based on points. In 2002, once a regulatory system was in place for the licensing of pension fund companies and a central clearinghouse had been established, it introduced a mandatory funded defined-contribution pension scheme and a voluntary scheme to supplement the two mandated schemes. Characteristics of Croatia's Pension System The main characteristics of the Croatian pension system include the design of the individual pillars of social insurance; the rules governing pension system taxation; and the institutional structure, coverage, and provisions governing old-age, disability, and survivorship pensions. The design of the pension system is assessed using a conceptual framework developed by the World Bank, which generally recommends including a funded component if conditions are appropriate but increasingly recog- nizes that a range of choices is available to policy makers to provide effec- tive old-age protection in a manner that is fiscally responsible (Holzmann and Hinz 2005). In general, the World Bank supports pension systems composed of some combination of five basic pillars: · a noncontributory (or zero) pillar (in the form of a demogrant, social pension, or social assistance benefit) intended to provide a minimal level of income protection 92 Adequacy of Retirement Income after Pension Reforms · a first-pillar contributory system linked to earnings, which seeks to replace a portion of preretirement income · a mandatory second pillar (essentially, individual savings accounts), which can be designed in various ways · a voluntary third pillar, which is flexible and discretionary (this pillar, too, can take a variety of forms) · a fourth pillar of informal intrafamily or intergenerational sources of financial and nonfinancial support to the elderly, including access to health care and housing. Pillar Design The reformed pension system provides old-age income support to the elderly through all five of these pillars (table 3.2.) The publicly managed noncontributory zero pillar, financed with general tax revenues, redistrib- utes income to lower income groups using means testing such that eligi- ble beneficiaries receive a benefit sufficient to provide them with a minimum state-defined income (which varies by household size). Benefits are adjusted on an ad hoc basis. Both the traditional publicly managed pay-as-you-go first pillar and the privately managed, fully funded second pillar are earnings-related schemes. Participation in both schemes is mandatory for new entrants and people who were under 40 when the reforms were implemented. First-pillar benefits are computed on the basis of a point system and indexed using a combination of wage and price growth (or Swiss indexation) whereby benefits increase with wages but at a lower rate. Second-pillar benefits are a function of an individual's contributions, investment earnings, and life expectancy at retirement. The third pillar is an optional privately managed, fully funded defined-contribution pension scheme intended to provide individuals with a mechanism for supplementing the benefits paid by the mandatory pillars. The fourth pillar provides health care to the elderly as part of the national health care system. First-pillar contributions are exempt from taxation, while benefits are taxed. The fully funded second and third pillars are subjected to exempt-exempt-taxed taxation (a classic expenditure tax), meaning that contributions are exempt from taxation and investment income is exempt but benefits are taxed (see box 1.1 in chapter 1). The noncon- tributory zero pillar (which provides a means-tested benefit for the poor) and the fourth pillar (which provides health care coverage) are completely tax exempt. Table 3.2 Structure of the Croatian Pension System Taxation Investment income/ capital Scheme type Coverage Type Function Financing Generic benefit Benefit indexation Contributions gains Benefits Zero pillar (public Universal Means tested Redistributive Tax revenues Certain percentage of Ad hoc n.a. n.a. Exempt noncontributory) state-defined benefit, depending on house- hold size First pillar (public, Mandatory Point Insurance Percentage of Depends on individual 50 percent prices, Exempt n.a. Taxed earnings related) individual earnings wage earned in relation 50 percent wages to average wage and length of coverage Second pillar Mandatory Defined Insurance Percentage of Annuity from capital Consumer price Exempt Exempt Taxed (private, earnings contribution individual earnings accumulation index related) Third pillar (private, Voluntary Defined Insurance Voluntary contributions Pension from capital Depends on options Exempta Exempt Taxed voluntary) contribution accumulation chosen Fourth pillar (public Mandatory n.a. Insurance Percentage of Specified health service n.a. Exempt n.a. n.a. health care) individual earnings package plus tax revenues Sources: Anusic 2007; Anusic, O'Keefe, and Madzarevic-Sujster 2003; INPRS 2003. n.a. = Not applicable. a. Contributions of up to 12,000 kunas per year can be deducted from personal income for tax purposes. 93 94 Adequacy of Retirement Income after Pension Reforms Noncontributory scheme. A noncontributory social assistance scheme provides financial support to households whose income falls below a min- imum threshold. The program is open to anyone, including the elderly. Benefits are means tested. The amount of the benefit is set as a percent- age of the state-defined subsistence allowance. The percentage depends on the applicant's age and the size of his or her household. Benefits are adjusted on an ad hoc basis. In 2005, 2.7 percent of the population received a social assistance benefit, at a cost equivalent to 0.22 percent of GDP (World Bank 2007a). Earnings-related schemes. Both the traditional, publicly managed, pay- as-you-go first pillar and the privately managed, fully funded second pil- lar are earnings-related schemes. The defined-benefit formula used to calculate pensions under the traditional first-pillar scheme was eliminated in 1998 in favor of a new formula based on points, which are determined by an individual's wages relative to the average wage (table 3.3). The number of years of wages on which benefits are based is gradually increas- ing, from 10 years to the entirety of an individual's service, thereby tight- ening the link between the individual's lifetime contributions and the benefits he or she receives in retirement. The fully funded second-pillar pension scheme was made optional for people age 40­50 but mandatory for everyone under age 40 at the time the scheme was introduced. Because low retirement ages (60 for men and 55 for women) were partly responsible for the fiscal imbalances of the pension system, the reforms raised retirement ages by six months a year, starting in 2000, such that the ages reached 65 for men and 60 for women in 2008. The fact that women can still retire five years earlier than men is problematic, given that women are expected to live eight years longer on average and to collect benefits for more than twice as long as men.2 To finance the two mandated schemes, employees contribute 20 percent of their wages, 5 percentage points of which go the funded second pillar. For individuals enrolled only in the first-pillar scheme (most of whom are older), all of their contributions go to the first pillar. Voluntary scheme. The voluntary third-pillar scheme was introduced in 2002, at the same time as the mandatory second-pillar scheme. All adult citizens may participate in the scheme, and employers can make contribu- tions on behalf of participating employees (table 3.4). To encourage par- ticipation, the government matches 25 percent of the contributions made Table 3.3 Parameters of Earnings-Related Schemes in Croatia before and after Reform Pension Vesting Contribution Contribution assessment Pillar Stage period rate ceiling Benefit rate base Retirement age First pillar Prereform 15 years 25.5 percent 2.2 percent for men and 10 best 55 for women and (earnings related) 2.5 percent accrual rate for consecutive 60 for men women years'wages Postreform 15 years 20 percent all 5 times the 0.75 percent a year for an Gradually Increasing from employeesb average wage average worker participating increasing to gradually to in the first pillar only; for full-career by 60 for women second-pillar participant, 2010 and 65 for men 0.25 percent of first-pillar by 2008 benefits plus 0.25 percent of average wage (flat) Second pillar Prereform n.a. n.a. n.a. n.a. n.a. n.a. (earnings related) Postreform 15 years 5 percent by 5 times the Life annuity for single people; Accumulated Increasing employee average wage mandatory joint- and-survivor funds gradually to annuity for married couple. If 60 for women both spouses (without and 65 for men children) agree, they can take by 2008 single annuities. People with children under age 18 receive a mandatory annuity with a guarantee period until child reaches age 18. Sources: Anusic, O'Keefe, and Madzarevic-Sujster 2003; consultations with World Bank staff. 95 n.a. = Not applicable. a. Until 2009, men are eligible for a pension after completing 40 years of service (for women, the requirement is 35 years) and reaching the retirement ages specified under the prereform scheme. b. This amount represents the total contribution rate for individuals participating in the first pillar only. Individuals participating in both the first and the second pillars pay 15 percent to the first pillar. 96 Adequacy of Retirement Income after Pension Reforms Table 3.4 Characteristics of the Voluntary Scheme in Croatia Contributions Lump-sum Vesting Retirement Tax advantages tax deductible payments possi- Coverage period age to participants by employers ble in retirement All citizens Noa 50 Yes No Yes Source: Anusic 2007. a. Benefits can be collected once an individual retires from the mandated schemes or upon reaching age 50. to an individual's account, up to an annual contribution ceiling of 5,000 kunas (HRK). The government further supports the scheme by allowing individuals to deduct contributions up to HRK 12,000 annually from their income for tax purposes. (Surprisingly, the government does not allow employers to deduct contributions made to the third pillar on behalf of employees. Most countries encourage employers to contribute to volun- tary schemes by offering some form of tax incentive.) Third-pillar benefits are taxed as regular income under the income tax law. Benefits can be collected once an individual retires from the mandated schemes or upon reaching age 50. They can be received in the form of an annuity, a sched- uled withdrawal, or a lump-sum payment (which cannot exceed 30 percent of the account balance). Funds cannot be withdrawn before an individual reaches age 50, except if the individual dies or becomes disabled. Funds accumulating in the individual accounts of the third-pillar scheme are invested by pension fund management companies. Pension funds can be open (funds that operate with no restrictions on membership) or closed (funds with restrictions on membership, typically open only to the employ- ees of one or more employers). In June 2008, 6 open pension funds were in operation, with 117,478 participants and HRK 727 million (0.3 percent of GDP) in assets, and 13 closed pension funds were in operation, with 15,000 participants and HRK 128 million (0.05 percent of GDP) in assets. Eight percent of the working-age population were enrolled in the scheme. The majority (63.7 percent) of third-pillar assets are invested in Croatian bonds and HRK deposits, despite investment restrictions that are liberal rel- ative to those applied to the funded second pillar. Health care system. Health care in Croatia is provided mainly through a mandatory insurance scheme administrated by the Croatian Institute for Health Insurance (HZZO), which is responsible for reimbursing covered health care expenditures as defined by law. As the primary purchaser of health care services in Croatia, HZZO also plays a key role in the process of defining and pricing services covered by the scheme. Croatia 97 Health care is financed primarily on the basis of contributions from the economically active population. Employers contribute 15 percent of employee wages to the scheme (employees pay nothing directly), plus an additional 0.5 percent of wages for occupational safety and workers com- pensation. Active duty military and people who are not economically active--including minors, students, the unemployed, the disabled, veter- ans, and the elderly--are not required to contribute. Contributions from 1.4 million economically active people covered roughly 80 percent of expenditures in 2002, with the remainder funded from transfers from the government. Given that the scheme provides health insurance coverage to 2.8 million people who do not contribute, the economically active popu- lation is subsidizing their coverage (Voncina 2006). Some 20 percent of beneficiaries of the health insurance scheme are required to make copay- ments for certain health care services and pharmaceuticals. Supplemental insurance is available for those seeking coverage for health care services not covered by the mandatory scheme and (since 2004) to pay for the cost of copayments for services provided under the mandatory scheme. Those elderly persons above the income threshold have subsidized flat premiums for supplementary insurance. In 2004, contributions for the supplemental scheme accounted for 3.5 percent of HZZO's total revenues (Voncina 2006). In 2005, health expenditures accounted for 7.4 percent of GDP, 81.3 percent of which was public expenditure and 18.7 of which was pri- vate. Of the private expenditures, 93.6 percent was attributable to out-of- pocket expenditures, in the form of informal payments, direct payments, and copayments (WHO 2008). Institutional Structure and Coverage of Earnings-Related Schemes Contributions for the first-pillar pension scheme are collected by the Croatian Tax Administration; the Croatian Pension Insurance Institute administers the scheme and is responsible for paying benefits. Responsi- bility for collecting all social contributions was transferred to the Croatian Tax Administration, which is responsible for collecting all other taxes, in order to increase administrative efficiency. Contributions for the second- pillar pension scheme are also collected by the Croatian Tax Administra- tion but are administered by the Central Registry of Insured People (REGOS), which also maintains the database of second-pillar contribu- tors and beneficiaries. Employers deposit first-pillar contributions with the Croatian Tax Administration's account and second-pillar contribu- tions with REGOS's account, both of which are maintained by the Croatian Treasury. The Agency for the Supervision of Financial Services 98 Adequacy of Retirement Income after Pension Reforms is responsible for overseeing REGOS and for licensing, monitoring, and supervising pension funds. In June 2008, four pension fund management companies were licensed to operate in the mandatory second-pillar pension scheme. They were serving 1.44 million contributors (54 percent of the working-age popula- tion and 82 percent of the labor force). Total participants in the scheme represented roughly 90 percent of the total insured population. In June 2008, assets of the scheme represented 8.2 percent of GDP. Structure of Benefits The earnings-related pension scheme provides old-age, disability, and sur- vivorship pensions. The provisions governing each of these types of ben- efits are discussed as follows. Old-age benefits. To claim an old-age pension, individuals must have at least 15 years of contributory service and have reached retirement age. Retirement ages have been rising at the rate of six months a year such that they will reach 65 for men and 60 for women in 2008. Men with 35 years of contributory service and women with 30 years of service may retire up to five years before reaching their retirement age, subject to a reduction in benefits of 1.8 percent a year of early retirement (before 2008, the reduction was 3.6 percent a year). Benefits for individuals enrolled in both the first- and the second-pillar schemes--everyone under age 40 when the reform was introduced and some older workers who were age 40­50 when the reform was introduced and who elected to join the second-pillar scheme--are paid from both schemes. First-pillar benefits (for those participating in both the first and the second pillars) for years of service realized in the first pillar only are calculated by multiplying the individual's points and the point value. For years of service realized in both of the two mandatory pillars, the basic pension is computed on the basis of two distinct components. The first component is earnings related and points based, with an individual's points determined by his or her wages relative to the average wage.3 This component is calculated by mul- tiplying total points earned after joining the second pillar by 25 percent of the point value. The point value is indexed on the basis of 50 percent inflation and 50 percent wage growth (Swiss indexation). The second component is a flat benefit computed by multiplying the individual's years of service in the reformed pension system and 0.25 percent of the average gross wage in the previous year.4 Benefits for individuals who are not participating in the second pillar are paid only from the first pillar and Croatia 99 are computed on the basis of the individual's years of contributory service and wages relative to the average wage in each year worked. Total points are calculated by multiplying the average points per year of service by the number of years of service. To calculate the benefit, total points earned are multiplied by the point value. Lifetime lower wage earners are eligi- ble for the minimum pension, which is calculated on the basis of an accrual rate of 0.825 percent applied to the average wage earned in 1998 and indexed on the basis of 50 percent inflation and 50 percent wage growth.5 Second-pillar benefits are a function of an individual's contribu- tions, investment earnings, and life expectancy at retirement.6 At the time of retirement, the capital accumulation is converted into life annuities. If workers are married, joint-and-survivor annuities are purchased. If both spouses (without children) agree, they can choose a single annuity. Disability benefits. Disability benefits are provided mainly by the first pillar of the reformed Croatian pension system. Individuals who lose some or all of their capacity to work are entitled to a disability pension (table 3.5). Depending on the degree of the incapacity, the individual will be entitled to a disability pension caused by either occupational or gen- eral incapacity. Occupational incapacity refers to someone with a perma- nent but partial disability (defined as having lost more than 50 percent of capacity but still capable of working); general incapacity refers to some- one who is incapable of working. Disability benefits depend on the degree of an individual's disability, the individual's wages relative to the average wage during the time he or she worked, and the length of contributory service to which service credit is awarded for years lost to disability. Table 3.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in Croatia Scheme Contribution Partial type Vesting period rate Eligibility Benefit rate pension First pillar Minimum coverage No specific Permanent loss 1 percent 80 percent (earnings of one-third of contribution in capacity for a per year for related) working life after rate for general disability an average age 20 (age 26 for disability pension; at least worker individuals with a benefits 50 percent loss university degree) in capacity for a partial disability pension Sources: Anusic, O'Keefe, and Madzarevic-Sujster 2003; U.S. Social Security Administration 2006. 100 Adequacy of Retirement Income after Pension Reforms Benefits are paid indefinitely, unless the individual's condition improves. If the total benefits the individual would have received from the first- and second-pillar schemes are lower than the disability pension to which he or she is entitled, the balance in the individual's second-pillar account is trans- ferred to the Croatian Pension Insurance Institute (the administrator of the first-pillar scheme), which pays the individual the higher disability pension. Survivor benefits. Survivor benefits are awarded to the dependents of individuals who, at the time of their death, were receiving (or had met the criteria to receive) an old-age or disability pension or had at least five years of contributory service (or 10 years of qualifying periods) (table 3.6). If total survivor benefits are higher than the benefits the deceased would have received in total from the first- and second-pillar schemes, the bal- ance in the deceased's second-pillar account is transferred to the Croatian Pension Insurance Institute. The value of survivor benefits is based on the number of the deceased's survivors and the old-age or disability pension to which the deceased was entitled. Survivor benefits are calculated on the basis of actual and imputed years of service (if the deceased died as a result of an occupational injury or disease) using a formula similar to that used to compute old-age pen- sions. In no case can the total value of benefits paid to all survivors exceed the benefit to which the deceased would have been entitled. Eligible survivors include spouses, orphans, parents, and siblings. Spouses who are age 45­49 when they become eligible survivors can begin collect- ing survivor benefits upon reaching age 50 for women (age 60 for men). Eligible spousal survivors must be caring for children or the disabled or have been fully dependent on the deceased. If a surviving spouse is younger than age 45, the spouse is eligible for only a one-year transition benefit. Orphans are eligible for benefits through age 15 (age 18 if unemployed, age 26 if a full-time student), or indefinitely in cases where the orphan is disabled. Assessment of the Performance of Croatia's Pension System The World Bank has established four principles for evaluating public pen- sion systems, which together should guide the process of pension reform (see Holzmann and Hinz 2005). Broadly speaking, these principles include the adequacy and security of benefits, the affordability of contri- butions, the sustainability of the system over time, and the robustness of the system in the face of demographic changes and macroeconomic Table 3.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in Croatia Spouse Orphan Scheme replacement replacement Total family type Eligibility rate Benefit duration Remarriage test Orphan age limit rate benefit First pillar Eligibility of the 70 percent if the For life, unless the The pension ceases 15 (18 if unem- 70 percent if the 70 percent for one (earnings deceased for an spouse is the spouse remarries; if the spouse ployed, 26 if orphan is the survivor; 80 per- related) old-age or a only survivor also for life upon remarries and is full-time student) only survivor cent for two disability pension, remarriage if younger than 50 survivors; 90 per- a minimum of spouse remarries unless disabled cent for three 5 years of but is older than survivors; coverage or 50 or disabled 100 percent for 10 years of four or more qualifying periods survivors by the deceased Sources: Anusic, O'Keefe, Madzarevic-Sujster 2003; U.S. Social Security Administration 2006. 101 102 Adequacy of Retirement Income after Pension Reforms shocks. This chapter focuses primarily on the adequacy of benefits and the financial sustainability of the earnings-related pension scheme. The remaining principles are mentioned only briefly. Adequacy is analyzed through the lens of net replacement rates. Financial sustainability is evaluated using projections of pension expenditure and revenues. Benefit Adequacy Replacement rates are a useful yardstick for measuring the adequacy of pension benefits, because they express benefits relative to preretirement earnings, thereby indicating the degree to which income is replaced when workers retire. Two variants are commonly used. Gross replacement rates compute income replacement as the ratio of benefits paid to pretax pre- retirement earnings. Net replacement rates compute income replacement as the ratio of benefits received (that is, after the payment of taxes and other levies, including contributions for social insurance) to posttax pre- retirement earnings. In general, net replacement rates are a more useful measure of benefit adequacy, because they capture the degree to which actual take-home pay is replaced when workers retire. The level of income replacement at retirement is not the only meas- ure of benefit adequacy. For the full assessment of benefit adequacy, it is also important to determine how postretirement indexation rules will affect replacement rates during retirement. Pension benefits in retire- ment are expected to be indexed to inflation, so that their real value is maintained. In a growing economy with rising real wages, however, mere price indexation of pensions leads to a deterioration of the relative con- sumption position of the retirees. Individuals with otherwise identical work histories will receive different pensions depending on when they retire. For this reason, some countries have introduced mixed indexation of pensions that use varying weights of inflation and wage growth in the indexation formula. For the evaluation of the effect of indexation on replacement rates in Croatia, the replacement rates are normalized to 100 and the assump- tions for calculating the replacement rates are maintained (that is, infla- tion is 2.5 percent a year and real wage growth is 2 percent a year). The change in the replacement rate is measured in comparison to full wage indexation or compared to an active worker (that is, in active earnings units). The results of this analysis indicate that the relative income posi- tion of a retiree would deteriorate by 12 percent after 10 years in retire- ment and by 36 percent after 35 years in retirement. The following evaluation of income replacement considers replacement rates only at Croatia 103 retirement and does not take into account the impact of indexation policies on replacement rates during retirement. Replacement rates are a function of the formula governing pension benefits; an individual's contribution history; and, in the case of net replacement rates, the rules of income tax, social security contributions, and other relevant levies. The benefit formula establishes the degree to which the system redistributes income across individuals of different lev- els of preretirement earnings. Progressive systems provide higher levels of income replacement to people with lower levels of preretirement income. In general, the degree to which a system is redistributive depends on the existence (and value) of flat transfers and minimum pension guarantees, the degree to which benefits are earnings related, and the existence of ceilings on earnings subject to contributions. An individual's contribution history can be characterized by his or her age of entry into the labor force, contribution density, and decisions regarding the timing of retirement. To some degree, these three factors are influenced by the incentives embod- ied in the pension system. The tax and contribution system influences net replacement rates through the progressiveness of the income tax formula, which taxes (higher) income during a worker's active life more than it taxes (lower) pension benefits in retirement. In addition, social security levies (for pensions; unemployment; health care; and, at times, housing and family benefits) are typically reduced or eliminated altogether in retirement. These benefits are particularly important for low- to middle- income groups. Benchmarks need to be established in order to evaluate the adequacy of the income replacement provided by the earnings-related pension schemes. Unfortunately, there is no consensus on what constitutes ade- quacy. According to one widely respected definition, pensions are ade- quate when they are sufficient to prevent poverty among the elderly and provide the vast majority of the population with a reliable mechanism for smoothing income over a lifetime. Even with a definition, however, establishing benchmarks is problematic, because attitudes vary across countries as a result of social and cultural perceptions. Moreover, bench- marks ignore the other factors affecting the welfare of the elderly--and varying across countries--including the existence and generosity of health insurance and long-term care, the cost of housing, the structure of traditional living arrangements, the presence of informal intrafamily or intergenerational sources of financial and nonfinancial support, and the availability and security of other mechanisms for people to save for their own retirement. 104 Adequacy of Retirement Income after Pension Reforms One reputable nine-country study (OECD 2001) observes that living standards are roughly comparable for people 10 years older than the normal retirement age and people 15 years younger than the normal retirement age when retirees have disposable income equal to roughly 80 percent of the disposable income of working-age people. In part, this is attributable to the fact that retirees have no work-related expenses (they do not have to commute or buy special clothing or uniforms, for example). This finding, however, does not imply that mandatory first-pillar pension schemes should actually target an 80 percent net replacement rate. To the contrary, in middle- and high-income countries, one can reasonably expect individ- uals to save for their own retirement--and the empirical evidence sug- gests that, in practice, they do so.7 There is also some evidence to suggest that the ratio between preretirement and postretirement income is some- what independent of the income replacement mandate of the public pen- sion system. Put simply, individuals tend to save more in countries with more modest mandates (and vice versa). Because Croatia has access to relatively well-developed financial mar- kets, it would seem reasonable to expect middle- and higher-income workers to save enough to finance at least 25 percent, if not closer to 50 percent, of this 80 percent income replacement target. Given this, three benchmarks are provided: a 40 percent net replacement rate (which implies that individuals would be expected to save enough to finance half of the total income-replacement target); a 60 percent net replacement rate (which implies that individuals would be expected to finance a quarter of the target); and an 80 percent net replacement rate (which implies that individuals, most of whom would be low-income earners, would not be expected to contribute anything toward the tar- get).8 In the following analysis, these benchmarks are used to evaluate the adequacy of benefits in Croatia compared with the average net replacement rate observed in 53 countries around the world, the average net replacement rate observed in selected countries in Europe and Central Asia, and the poverty line in Croatia.9 To estimate gross and net replacement rates, we use the Analysis of Pension Entitlements across Countries (APEX) model to consider two critical dimensions: earnings levels and contribution periods.10 This model generates estimates for replacement rates under steady-state assumptions (that is, as if the rules of the reformed pension scheme had been in place over the entire active life of the individual). Because life expectancies at retirement are projected to increase over time--which will affect the benefits paid by defined-contribution pension schemes--a reference year Croatia 105 must be chosen. For the purpose of this study, 2040 is used, because it provides a sufficiently long contribution period over which to approxi- mate steady-state conditions. The first critical task is to investigate levels of income replacement across a relevant spectrum of income. Income is represented as a percent- age (50­200 percent) of average earnings. The second task is to investi- gate the impact on income replacement of differences in the duration, timing, and density of an individual's contribution history (density refers to the percentage of time an individual actually contributes over a given period). For the facilitation of the presentation of these multidimensional results, replacement rates are computed as a function of the age an indi- vidual exits the labor market. They are presented separately for full-career and partial-career workers. Replacement rates for full-career workers. Projected replacement rates for full-career workers in 2040 are examined first. For the purpose of this analysis, a full-career is defined as continuous employment from age 20 to age 65. Gross replacement rates are presented in figure 3.2 as a func- tion of an individual's preretirement income relative to the economywide average wage. This figure reveals the degree to which income is redistrib- uted under the Croatian pension system as a result of the existence of a Figure 3.2 Sources of Gross Replacement Rates in Croatia, by Income Level 60 50 (percent) 40 rate 30 20 replacement 10 gross 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (points system) flat benefit Source: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. 106 Adequacy of Retirement Income after Pension Reforms flat benefit, which contributes less to total benefits as income rises. The shares of income replacement attributable to the points-based first-pillar benefit and to the second pillar are constant across income. The situation does not change significantly when taxes are taken into consideration (figure 3.3). The effect of taxes and contributions on replace- ment rates increases as the level of income rises. The tax system does not affect the overall picture of replacement rates. As is the case with gross replacement rates, net replacement rates fall with income. Net replacement rates by income level are presented in figure 3.4 (in which net replacement rates are represented by the dotted line).11 Preretirement income is expressed relative to the economywide average wage. The three benchmarks discussed previously are represented by three short horizontal lines abutting the y-axis. The three downward sloping lines represent the world average, the regional average, and the Croatian average,12 where each of these indicators is expressed relative to the econ- omywide average wage. The figure suggests that pensions for middle- and high-income full- career workers in Croatia can generally be considered adequate. Replacement rates for middle-income workers are around the 60 percent Figure 3.3 Sources of Net Replacement Rates in Croatia, by Income Level 100 90 80 (percent) 70 rate 60 50 40 replacement 30 of 20 10 share 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (points system) taxes flat benefit Source: APEX model. Note: Figure shows projected replacement rate for full-career worker in 2040 as approximation of steady-state conditions. Croatia 107 Figure 3.4 Net Replacement Rates for Male Full-Career Workers in 2040 in Croatia, Europe and Central Asia, and the World 100 high benchmark 90 80 70 (percent) 60 middle benchmark rate 50 40 low benchmark 30 replacement 20 net 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage world average Europe and Central Asia average Croatia average poverty line (percentage of average income) Source: Authors' calculations based on World Bank 2007b and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. benchmark while high-income workers receive replacement rates only slightly below the 60 percent benchmark. This suggests that the pension system is effectively smoothing consumption from work into retirement for these workers. In contrast, replacement rates for low-income workers are about 10­15 percentage points below the 80 percent benchmark despite the existence of a flat benefit component. Given that benefits for all income levels exceed the poverty line, the objective of poverty alleviation is being met. The degree of redistribution in the Croatian pension system is roughly comparable to the regional and world averages (in all cases, net income replacement falls as preretirement income rises), but the levels of income replacement are lower than those provided in some other countries. Replacement rates for partial-career workers. Not everyone works from age 20 to the statutory retirement age. Many individuals enter and exit the labor force (often at different ages and for different periods of time) and earn different wages while working. To examine the adequacy of benefits 108 Adequacy of Retirement Income after Pension Reforms Figure 3.5 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Croatia, by Career Type and Exit Age 120 100 high benchmark 80 (percent) world and Europe and Central Asia average rate middle benchmark 60 low benchmark 40 replacement poverty line (percentage of 20 average income) net 0 45 50 55 60 65 70 exit age from labor market career type A: career type B: career type C: entry age--25, entry age--30, entry age--25, contribution contribution contribution density--100% density--100% density--75% Source: Authors' calculations based on World Bank 2007b and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. See text for description of career types. for partial-career workers, we consider three stylized cases (figure 3.5).13 These cases include career type A (someone entering the labor force at age 25 who works continuously for a period of years before leaving the workforce at some point between the ages of 50 and 70 and then claims a benefit); career type B (identical to career type A, except that the worker enters the workforce at age 30 and leaves no earlier than age 55); and career type C (identical to career type A, except that the individual contributes in only three years out of four while in the labor force).14 In cases in which the withdrawal from the formal labor market occurs before the statutory retirement age, the pension is claimed (and the replacement rate calculated) only at the later age. For withdrawals after the statutory retirement age, the ages coincide. Several conclusions can be drawn from this figure. First, some partial- career workers receive levels of income replacement lower than the poverty line, especially those retiring at very young ages. Second, leaving the workforce very early can be very costly. Someone retiring long before reaching the retirement age may not receive levels of income replacement Croatia 109 higher than even the lowest of the three benchmarks.15 Third, entering the workforce later in life is costly. Someone entering the workforce at age 30 receives a net replacement rate up to nine percentage points lower than someone entering the workforce at age 25. Fourth, working intermittently is costly. Someone who enters the workforce at the same age but who contributes only three years out of four will receive a net replacement rate that is 6­19 percentage points lower than someone who contributes continuously. In all cases, net replacement rates grow the longer someone continues to work. This is encouraging because it demonstrates that the pension system provides incentives for people to remain in the workforce. Career type A and B workers will attain the lowest of the three benchmarks before reaching the retirement age, while career type C workers must work to the retirement age. Similarly, career type A and B workers can attain the middle benchmark, provided that they work three to five years after reaching retirement age, but career type C workers will not be able to attain this benchmark, even if they work until age 70. Fiscal Sustainability The sustainability of a pay-as-you-go first-pillar pension scheme is best evaluated in actuarial terms by estimating the scheme's actuarial deficit as the difference between its assets and liabilities. If an actuarial deficit exists and is large, the scheme is financially unsustainable and needs policy actions that positively impact either its assets or liabilities, or both. A good proxy for the actuarial deficit is the difference between the present value of the scheme's expected future revenues (that is, contributions and other income) and expected future expenditures (that is, benefit payments, administrative costs, and other expenses) over an extended projection period. The difference between these two values represents an unfunded liability (sometimes referred to as a financing gap) on the public-sector balance sheet. Because this study is concerned also with the time path of revenues and expenditures (and the resulting balance across the projection period ending in 2040), this more pragmatic approach has been taken, and projections of expenditures, revenues, and deficits are presented on the basis of available postreform fiscal projections. As a result of the government's reforms, revenues are now projected to remain stable while expenditures will drop gradually, from 12 per- cent of GDP in 2005 to about 9 percent of GDP by 2040, such that the deficits of the pension scheme will grow progressively smaller by 2040 (figure 3.6). 110 Adequacy of Retirement Income after Pension Reforms Figure 3.6 Projected Fiscal Balance of Croatia's Public Pension System after Reform, 2005­40 14 12 10 GDP 8 of 6 4 2 percentage 0 ­2 ­4 2005 2010 2015 2020 2025 2030 2035 2040 year expenditures revenues deficit Source: Anusic 2007. Conclusions To address growing deficits in the pension system and projections that demonstrated that the aging of the population would place increasing pres- sure on the finances of the system over time, Croatia began a process of pension reform in the mid-1990s. Three years later, in 1998, it eliminated the traditional defined-benefit formula used to calculate pensions in favor of a new formula based on points and a flat benefit to effect income redis- tribution and protect low-income workers. In 2002, once a regulatory sys- tem was in place for the licensing of pension fund companies and a central clearing house had been established, the government introduced a manda- tory funded second-pillar scheme and a voluntary third-pillar scheme to supplement the two mandated schemes. Together, the reforms tightened the link between contributions and benefits and put the pension system on track to eventually becoming financially self-supporting by about 2040. The resulting projected gross and net replacement rates for full-career workers are roughly in line with regional and world averages, although they tend toward the lower end of the range (see chapter 1). Net replace- ment rates for full-career workers are projected to be 57­70 percent across the analyzed income spectrum. The stereotypical full-career worker is not, of course, representative of the average worker in Croatia. As is the case in other countries, workers with less than full careers--because they left the workforce before reaching retirement age, worked intermittently, or Croatia 111 have gaps in their employment history--risk receiving income replacement that is closer to--or even below--the lower benchmark of 40 percent.16 Some workers, particularly those who enter the workforce later in life or who have intermittent work histories, may have to remain in the work- force well beyond the normal retirement age to attain sufficient levels of income replacement. Moreover, all workers, including full-career workers, may need to save outside the mandated schemes if they want to replace 80 percent of their preretirement income. To increase income replacement by one percentage point, a typical full- career worker would need to save about 0.50 percent of his or her earn- ings from age 40 to the current retirement age.17 Given that the voluntary third-pillar pension scheme remains small--despite the fact that the gov- ernment provides matching contributions and exempts contributions from taxable income (both provisions are subject to some restrictions)--an opportunity exists for deepening its reach, possibly by providing incentives for employers to contribute on behalf of their employees. Such incentives do not now exist in Croatia. Replacement rates have been computed under the assumption that funded pension schemes earn a rate of return of 1.5 percentage points more than wage growth. This earnings differential broadly reflects the performance of pension funds in OECD countries over the past 30 years. The earnings differential in emerging economies is almost twice as large (see Holzmann 2009). The performance of pension funds in Croatia since their inception, however, has been well below this benchmark. If such performance continues, the Croatian pension system will not be capable of delivering the replacement rates projected for 2040. This concern calls for a review of pension fund performance and accelerated progress in financial market development. Achieving a higher target for income replacement can be accom- plished by postponing retirement (either in lieu of, or in combination with, increasing savings through such mechanisms as voluntary pension schemes). Increasing the retirement age in step with increases in life expectancy at retirement is a natural choice, for both individuals and pol- icy makers, but it requires cross-sectoral policy reforms to enable elderly workers to continue to participate in the labor market.18 Notes 1. For more information regarding these population projections, see Reiterer (2008). 112 Adequacy of Retirement Income after Pension Reforms 2. Life expectancy at retirement is currently 73 years for men and 81 years for women, which implies that, on average, men will live 8 years and women 21 years after reaching retirement age (see Anusic 2007). 3. For the length of coverage before 2002, the benefit from the first pillar is calculated on the basis of points earned using the same formula used for individuals participating in the first pillar only (that is, the points earned for years of service before 2002 are multiplied by the point value). 4. The monthly average point value in 2008 is HRK 54.11. In January 2008, the average gross wage was HRK 7,357, and the net wage was HRK 5,019. As a result, the point value is 1.1 percent of net wages and 0.75 percent of gross wages, which translates into accrual rates for an average earner in a traditional benefit formula. 5. Average gross earnings in 1998 were roughly HRK 6,000. 6. Benefits from second- and third-pillar schemes are typically lower than the actuarially neutral benefits presented in these scenarios, because pension funds typically assume life expectancies that are higher than the expectancies derived from general statistics for the entire population in order to protect themselves from adverse selection and other risks. 7. In Chile, for instance, 70 percent of retirees from the mandatory public pen- sion system own their home, which is a form of savings (see Valdés-Prieto 2008). 8. These benchmarks approximate the standards developed by the International Labour Organization (ILO) (1952) and by the Council of Europe (1990). ILO Convention 102 of 1952 sets a minimum benefit equal to 40 percent of the reference wage for married men of pensionable age. This amount was raised to 45 percent in 1968. The European Code of Security of 1990 sets a minimum standard for members of the Council of Europe equal to 65 per- cent for married people of a specific age. 9. As a proxy for the poverty line, this study uses 35 percent of the average net wage, which very broadly approximates a US$2.25-a-day poverty line con- verted into national currency, adjusted for purchasing power parity, expressed relative to the national average net wage, and averaged across the nine study countries. Such an approach enables valid comparisons to be made across the sample (see chapter 1). 10. The APEX model was developed by Axia Economics, with funding from the OECD and the World Bank. The model codes detailed eligibility and benefit rules for first- and second-pillar schemes based on available public informa- tion that has been verified by country contacts. Because the details of the rules sometimes change on short notice (and limited public disclosure), the calcu- lations presented here should be considered as best approximations only. 11. These are simulated replacement rates for an unmarried male working a hypo- thetical career path under the assumption that real wage growth is 2 percent, Croatia 113 inflation is 2.5 percent, the rate of return on invested assets is 3.5 percent, and the worker retires at the statutory retirement age. 12. As a proxy for the poverty line, this study uses 35 percent of the average net wage because this percentage very broadly approximates a US$2.25-a-day poverty line converted into national currency, adjusted for purchasing power parity, expressed relative to the national average net wage, and averaged across the nine study countries. Such an approach enables valid comparisons to be made across the sample. See chapter 1. 13. Only middle-income partial-career workers are examined because replace- ment rates are roughly comparable for workers with lower or higher levels of preretirement income. 14. Only middle-income partial-career workers are examined because replace- ment rates are roughly comparable for workers with lower or higher levels of preretirement income. 15. The drop in replacement rates for Category Type A workers is attributable to the actuarial reduction for early retirement. Workers with enough years of service (35 years) for early retirement can retire at age 60 if they so choose, but their benefits are subjected to penalties for early retirement. After age 60, replacement rates start increasing because, in addition to receiving more ben- efits for each year of additional service, early retirement is also penalized less as age retirement increases from age 60 to age 65. 16. The average effective years of service fell from 33 in 1999 to 29 in 2007. For 30 years of service and given the assumptions used in the analysis, the APEX model generates a replacement rate of 45 percent. 17. This estimate is based on the assumption that real wage growth is 2 percent, the net real rate of return on invested assets is 3.5 percent, and benefits (both from the unfunded and the funded pillars) are price indexed. 18. See Holzmann, MacKellar, and Repansek (2009) for a conference volume that addresses theses issues for the countries in southeastern Europe. Bibliography Anusic, Z. 2007. "Pension System and Pension Reform in Croatia." Background paper prepared for Public Finance Review, World Bank, Washington, DC. Anusic, Z., P. O'Keefe, and S. Madzarevic-Sujster. 2003. Pension Reform in Croatia. Washington, DC: World Bank. Council of Europe. 1990. European Code of Social Security (Revised). Rome. Holzmann, R., ed. 2009. Aging Populations, Pension Funds, and Financial Markets: Regional Perspectives and Global Challenges for Central, Eastern, and Southern Europe. Washington, DC: World Bank. Holzmann, R., and R. Hinz. 2005. Old-Age Income Support in the 21st Century. Washington, DC: World Bank. 114 Adequacy of Retirement Income after Pension Reforms Holzmann, R., L. MacKellar, and J. Repansek, eds. 2009. Pension Reform in South- eastern Europe: Linking to Labor and Financial Market Reforms. Washington, DC: World Bank. INPRS (International Network of Pension Regulators and Supervisors). 2003. Complementary and Private Pensions throughout the World. Geneva: INPRS. ILO (International Labour Organization). 1952. ILO Convention 102. Geneva: ILO. ------. 1967. ILO Convention 128. Geneva: ILO. OECD (Organisation for Economic Co-operation and Development). 2001. Ageing and Income: Financial Resources and Retirement in 9 OECD Countries. Paris: OECD. Reiterer, A. 2008. "Population Development and Age Structure in Southeastern Europe until 2050." World Bank, Washington, DC. U.S. Social Security Administration. 2006 Social Security Systems throughout the World: Europe. Washington, DC: Social Security Administration. Valdés-Prieto, S. 2008. Designs for the First-Pillar Pensions and the 2008 Chilean Reform. http://editorialexpress.com/cgi-bin/conference/download.cgi?db_ name=SECHI2008&paper_id=130. Voncina, L. 2006. Health Systems in Transition: Croatia. Geneva: World Health Organization. World Bank. 2005. Growth, Poverty, and Inequality: Eastern Europe and the Former Soviet Union. Washington, DC: World Bank. ------. 2007a. Croatia Living Standards Assessment. Washington, DC: World Bank. ------. 2007b. Pensions Panorama. Washington, DC: World Bank. WHO (World Health Organization). 2008. Database. http://www.who.int/ research/en/. C H A P T E R 4 The Czech Republic The Czech Republic inherited from the former Czechoslovakia a public pension system financed on a pay-as-you-go basis (meaning that contri- butions from current workers are used to pay benefits to current benefi- ciaries). Within a few years of the country's transition to a market economy, the fiscal balance of the pension system began to deteriorate. By 1997, the system was generating an annual deficit equivalent to 0.5 percent of gross domestic product (GDP). Pension expenditures con- tinued to increase while revenues remained constant. As a result, by 2000 the deficit had doubled to 0.9 percent of GDP. Recognizing that these deficits were not sustainable and that the sys- tem would face even greater challenges in the medium to long term as the population ages, the government began to undertake a set of paramet- ric reforms starting in the mid-1990s. These reforms--which included gradually increasing the retirement age and changing the benefit formula and indexation rules--improved the fiscal balance of the system.1 In 1994, the government introduced a voluntary funded pension scheme that ben- efits from direct government subsidies. In July 2008, parliament approved additional parametric changes. Despite the parametric reforms of 1995 and 2003, the Czech pen- sion system remains insufficiently prepared for looming demographic 115 116 Adequacy of Retirement Income after Pension Reforms change, because it is based predominantly on a single pay-as-you-go, defined-benefit pension scheme. By 2050, the old-age dependency ratio (the population age 65 and older divided by the population age 20­64) is projected to increase to roughly 60 percent, up from 21.7 percent in 2005. Over the same period and before the recently announced reform, pension system revenues were projected to level off at 8.1 percent of GDP while expenditures were projected to rise to 11.7 percent of GPD, resulting in a deficit of 3.4 percent of GDP. The 2008 reform is estimated to reduce expenditures by 1.2 percent of GDP by 2050 and to ensure financial sus- tainability until 2030, after which a deficit will emerge, eventually reaching 2.5 percent of GDP by 2050. This suggests that further reforms will be required to achieve both adequate and sustainable pensions in the future. Against this backdrop of demographic change, this chapter evaluates the Czech pension system, focusing on fiscal sustainability and benefit adequacy. Adequacy is evaluated through the lens of statutory net replace- ment rates for different retirement ages, patterns of contributions, and income levels with comparisons to international benchmarks. The chapter is organized as follows. The next section discusses the motivation for the reforms. The following section describes the key char- acteristics of the reformed pension system. The third section assesses the adequacy of pension benefits and the fiscal sustainability of the system. The last section draws conclusions. Motivation for Reform The Czech Republic inherited a socialist-era pension system that suffered from a number of serious design flaws. Benefits were comparatively gener- ous, particularly for certain occupations. Lax disability criteria resulted in more disability beneficiaries than were justified on medical grounds. The redistributive benefit formula, augmented by a flat state-compensation benefit introduced in 1990, weakened the link between contributions and benefits. The retirement age was low relative to life expectancy, a problem exacerbated by the generosity of benefits (which increased incentives for early retirement). At the individual level, this increased the value of retire- ment benefits relative to contributions. For the pension system, this increased the total number of beneficiaries relative to contributors. The impact of these factors--together with a decline in GDP and increasing unemployment (albeit at levels not nearly as high as in many other countries in the region)--undermined the fiscal health of the pen- sion system. In 1995, two years before the system actually began running The Czech Republic 117 Table 4.1 Fiscal Balance of the Czech Republic's Pension System before Reform, 1994­2000 (percentage of GDP) Year Revenues Expenditures Balance 1994 8.3 7.1 1.2 1995 8.2 7.5 0.7 1996 8.1 7.8 0.3 1997 8.4 8.8 ­0.4 1998 8.3 9.0 ­0.7 1999 8.4 9.4 ­0.7 2000 8.6 9.5 ­0.9 Source: Lasagabaster, Rocha, and Wiese 2002. deficits (table 4.1), policy makers implemented a set of parametric reforms intended to restore long-term sustainability. These reforms included gradually increasing the retirement age, eliminating special priv- ileges for certain occupations, tightening eligibility conditions for disabil- ity benefits, and introducing a benefit formula based on both a flat-rate and an earnings-related benefit. In 2001, penalties for early retirement were increased to levels closer to actuarial neutrality in order to contain the rising number of early retirees. Characteristics of the Czech Republic's Pension System This section describes the main characteristics of the Czech Republic's pension system. These characteristics include the design of the individual pillars of social insurance; the rules governing pension system taxation, institutional structure, and coverage; and the provisions governing old-age, disability, and survivorship pensions. The design of the pension system is assessed using a conceptual framework developed by the World Bank, which generally recommends including a funded component if conditions are appropriate but increasingly recognizes that a range of choices is avail- able to policy makers to provide effective old-age protection in a manner that is fiscally responsible (see Holzmann and Hinz 2005). In general, the World Bank supports pension systems composed of some combination of five basic pillars: · a noncontributory (or zero) pillar (in the form of a demogrant, social pension, or social assistance benefit) intended to provide a minimal level of income protection; · a first-pillar contributory system linked to earnings, which seeks to replace a portion of preretirement income; 118 Adequacy of Retirement Income after Pension Reforms · a mandatory second pillar (essentially, individual savings accounts), which can be designed in various ways; · a voluntary third pillar, which is flexible and discretionary (this pillar, too, can take a variety of forms); and · a fourth pillar of informal intrafamily or intergenerational sources of financial and nonfinancial support to the elderly, including access to health care and housing. Pillar Design The design of the Czech Republic's pension system incorporates four of the five pillars recommended by the World Bank (table 4.2). The pub- licly managed noncontributory zero pillar, financed with general tax revenues, redistributes income to lower income groups using means testing such that eligible beneficiaries receive a benefit sufficient to ensure them a total income equal to the state-defined minimum income guarantee. Benefits are adjusted for inflation, with the objective of ensuring their real value over time. The mandatory first pillar, financed on a pay-as-you-go basis, provides a flat benefit plus an earnings-related benefit. The flat benefit is awarded to all pensioners, irrespective of pre- retirement income or length of service. Because the earnings-related benefit is also progressive, the first pillar is highly redistributive. There is no second pillar (that is, a mandatory funded component). The vol- untary third pillar, introduced in 1993 to supplement the benefits of the first pillar, is a defined-contribution scheme in which a partici- pant's benefits depend on his or her contributions and accumulated investment earnings to the point of retirement. The mandatory fourth pillar, financed by a combination of contributions and general tax rev- enues, provides health insurance to the elderly as part of the national health care system. The third pillar is subjected to an exempt-exempt-taxed (EET) regime, meaning that contributions are partially exempt from taxa- tion, investment income is fully exempt, and benefits are taxed (see box 1.1 in chapter 1). This is similar to the tax regimes of most of the countries in the Organisation for Economic Co-operation and Development (OECD), which take a more classical EET approach. Benefits paid by the zero pillar are exempt from taxation, contribu- tions to the first pillar are taxed, and benefits are partially exempt from taxation. Contributions to the health care system (the fourth pillar) are taxed. Table 4.2 Structure of the Czech Republic's Pension System Taxation Investment income/ Benefit capital Scheme type Coverage Type Function Financing Generic benefit indexation Contributions gains Benefits Zero pillar Universal Means tested Redistributive Tax revenues Difference between Prices n.a. n.a. Exempt (public minimum defined- noncontributory) benefit level and actual income First pillar Mandatory Defined benefit Insurance Percentage of Flat benefit plus A minimum infla- Taxeda n.a. Exemptb (private, individual earnings-related tion rate of plus earnings related) earnings pension one-third of real wage growth Third pillar Voluntary Defined Insurance Voluntary Pension from Depends on Exemptc Exempt Taxed (private, contribution contributions capital options voluntary) accumulation chosen Fourth pillar Mandatory n.a. Insurance Percentage of Specified health n.a. Taxed n.a. n.a. (public individual service package health care) earnings plus tax revenues Source: Authors' compilation based on data from European Commission 2007a, 2007b; Hemmings and Whitehouse 2006; and Rokosova and Schreyogg 2005. n.a. = Not applicable. a. Since January 1, 2008, contributions paid to the first pillar are part of the income tax base, consistent with the tax reform that introduced a flat-rate tax of 15 percent. b. Pensioners are provided with a large tax allowance on pension income. Pensioners with total taxable income of less than 80 percent of average earnings do not pay income tax, which 119 effectively exempts pensions from taxation for all pensioners except those with substantial income from voluntary pension schemes or other sources. Pensioners who are liable for taxes pay a rate lower than that levied on earnings. As a result, they pay substantially less in taxes than do workers with the same total income (Hemmings and Whitehouse 2006). c. Contributions of 6,000 koruny­12,000 koruny and employer contributions up to 5 percent of wage are exempt from taxation. 120 Adequacy of Retirement Income after Pension Reforms Noncontributory scheme. The elderly in the Czech Republic are eligible for means-tested noncontributory benefits (living minimum) as part of a social assistance scheme intended to guarantee, without any categorical bias, a minimum level of income to the overall population. To be eligible, applicants need to demonstrate that their total income--from gainful activity, revenues from investments, and other regular activity (net of taxes and social insurance contributions)--falls below the state-defined minimum. Eligibility also requires demonstration of a willingness to work (except for those below age 18 or above age 65), strengthening the incen- tives to seek employment. The minimum is computed as the sum of two parts.2 The first part is based on individual income needs (which are determined on the basis of the individual's age). The second part is based on household income needs (which are a function of the number of household members). In 2007 and 2008, a single-person household received a total benefit of 3,126 koruny (CZK). About 4 percent of the households in the Czech Republic are registered beneficiaries of the social assistance system. Earnings-related scheme. Workers in the Czech Republic are enrolled in a first-pillar earnings-related defined-benefit public pension scheme financed on a pay-as-you-go basis. To address rising fiscal deficits-- attributable in part to lax disability eligibility conditions and special privileges awarded to certain groups--and prepare for the aging of the population, the government reformed the parameters of the scheme in 1995, 2003, and 2008 (table 4.3). In January 1996, retirement ages were raised at the rate of two months per year for men and four months per year for women. In January 2004, an additional gradual increase in retire- ment ages was introduced, equalizing ages for men and childless women at 63 years and setting the age for other women at 59­62 years, based on the number of children. The 2008 reform, which will be implemented starting in 2010, raised the ages further, such that the normal retirement age will--between 2017 and 2030--reach 65 years for men and women with one or no children and 62­64 years for women with more than one child. At the same time, opportunities for early retirement in the future have been increased, albeit with some actuarially inspired reductions (dis- cussed later). The benefit formula is highly redistributive and includes a flat benefit (applied to all pensioners, regardless of service length or preretirement earnings) and an earnings-related benefit.3 Individuals with 25 years of service (by 2030, this will rise to 30 years) qualify for a full pension, with Table 4.3 Parameters of the First-Pillar Earnings-Related Scheme in the Czech Republic before and after Reform Contribution Pension Period Vesting period Contribution rate ceiling Benefit rate assessment base Retirement age Prereform n.a. n.a. n.a. n.a. Last 10 years' 60 for men, 53­57 for average women depending on earnings number of children valorized by wage growth Postreform 35 years at normal 28 percent None Flat benefit Gradually Gradually increasing to 65 retirement age by 2019; (21.5 percent by plus 1.5 increasing to by 2030 for men and for 20 years for people who employer, percent average of last women (with 0­1 retire five or more years 6.5 percent by per year 30 years of children), 64 (for women after the normal employee) earnings by with two children), 63 retirement age 2016 (for women with three children), and 62 (for women with four or more children) Sources: European Commission 2007a, 2007b; information provided by the Ministry of Labor and Social Affairs in July 2008. n.a. = Not applicable. 121 122 Adequacy of Retirement Income after Pension Reforms 1.5 percent of earnings credited for each year of service, not subject to any sort of a ceiling. The computation of the pension assessment base (that is, the wages used in the computation of benefits) is progressive, providing higher income replacement for lower wages with thresholds that are not subject to statutory indexation.4 The 1995 reforms also intro- duced changes to the benefit formula, including gradually increasing the number of years used in the pension computation from the past 10 years of wages to 30 years by 2016.5 Shifting toward the calculation of benefits based on lifetime earnings conforms with international practice and--to some extent--helps tighten the link between contributions and benefits. The degree to which the benefit formula is redistributive dilutes the ben- efits of this shift, however. Pensions are adjusted such that the average pension rises by at least the sum of inflation plus a third of the growth in real wages. The exact amount, which is set by decree, can be greater than the minimum stipulated under the law. The scheme is financed on a pay- as-you-go basis, with a total levy of 28 percent of earnings, of which 21.5 percentage points are paid by employers and 6.5 percentage points are paid by employees. Voluntary scheme. The Czech Republic introduced a voluntary pension scheme in 1994 (table 4.4). All citizens of the European Union (EU) age 18 and older who are also participating in either the first-pillar (earnings- related) public pension scheme or the public health-insurance scheme are eligible to enroll in the voluntary pension scheme. To proactively accommodate expected future changes in the first-pillar scheme-- changes that will likely reduce benefits--the government is actively encouraging participation in the voluntary scheme through a combina- tion of matching subsidies and tax exemptions. Contributions can be made by workers or their employers. The minimum monthly contribu- tion to the scheme is CZK 100; the minimum monthly state subsidy corresponding to that contribution is CZK 50 (equivalent to 50 percent of the actual contribution). Although the subsidy grows with the amount of the contribution, the ratio between the two gradually falls, such that a maximum subsidy of CZK 150 is provided for contributions of CZK 500 and higher. Annual contributions that exceed CZK 6,000 a year are deductible from personal income tax, subject to an annual maximum of CZK 12,000. The combination of state subsidies (in the form of matching contribu- tions) and tax exemptions implies that the government supports the first CZK 18,000 of individual voluntary pension savings--equivalent to about The Czech Republic 123 Table 4.4 Characteristics of the Voluntary Scheme in the Czech Republic Tax Contributions Lump-sum advantages tax payments Vesting Retirement to deductible possible in Coverage period age participants by employers retirement EU citizens age 18 and Five years 60 Yes Yes Yes older enrolled in the public pension or health insurance scheme Sources: Hemmings and Whitehouse 2006; Iglesias 2003. 8 percent of average earnings. Voluntary pension contributions made by employers are exempt from taxation up to 5 percent of gross earnings. Benefits from the voluntary pension scheme are taxed differently on the basis of whether they originate from employee contributions, state- provided matching contributions, employer contributions, or returns on invested assets. Taxes are levied on investment income at the rate of 15 percent; employer contributions are subjected to standard income tax provisions (Hemmings and White 2006). Supervision of the voluntary pension scheme is the responsibility of the Czech National Bank; the Ministry of Finance is responsible for over- seeing state-provided matching contributions. The assets of the voluntary pension scheme are invested by private management companies. There are currently 11 such companies. Under current law, management com- panies have some freedom to define the terms of the plans they offer, including options for receiving annuities and lump-sum payments.6 All plans must offer an old-age pension; disability pensions and survivor benefits are optional. Management companies are required to guarantee a 0 percent nominal rate of return (that is, participants are ensured of get- ting their money back, albeit without compensation for inflation). Participants may be enrolled in more than one pension plan, but they are restricted to making contributions to one plan at a time. Given the existence of state-provided matching contributions and preferential tax treatment (Iglesias 2003), participation in the voluntary scheme has expanded rapidly, from 3.5 percent of the labor force in 1994 to 57 percent by 2004.7 Older workers are more likely to participate than younger workers. Contributions have averaged roughly 2 percent of the average wage since 1999. The accumulated assets of the scheme were equivalent to 3.7 percent of GDP in 2004. 124 Adequacy of Retirement Income after Pension Reforms Health care system. Health care in the Czech Republic is provided mainly through a mandatory insurance scheme administrated by nine health insurance funds. In 2005, health expenditures accounted for 7.1 percent of GDP, 88.6 percent of which was public expenditure and the remaining 11.4 was private expenditure. Of private expenditure, 95.3 percent was attributable to out-of-pocket expenditures (informal payments, direct pay- ments, and copayments) (WHO database 2008). Participants have the right to choose their health insurance fund. Of the nine funds, the General Health Insurance Fund is, by far, the largest, covering 68 percent of the overall population in 2002 and the majority of the population for which the Ministry of Finance pays the premiums. Virtually all medical services are covered. These services include pre- ventive services, diagnostic procedures, ambulatory and hospital curative care, rehabilitation and care for the chronically ill, drugs and medical devices, medical transportation services, and spa therapy (when pre- scribed by a physician). For all medical services, the least-expensive avail- able treatment is fully covered. Only generic pharmaceuticals are fully covered; patients bear the cost of nongeneric drugs. Nongeneric drugs may be approved for reimbursement if there are no alternatives. A small num- ber of services are excluded (cosmetic surgery for nonmedical reasons and certain services made at a patient's request); a small number require copayments (certain kinds of dental care, including dentures). Prostheses, eyeglasses, and hearing aids may be partially or fully reimbursed. The cost of social care is not covered by the scheme but is, instead, borne directly by patients and the Ministry of Social Affairs. Within the limits established in the benefit package, the scheme provides free health care for all elderly people in all public facilities, irrespective of whether they made insurance premium contributions while they were working. Under the law, the premium for health insurance is 13.5 percent of pretax wages, subject to a floor of 25 percent of the national average wage and a ceiling of six times the national average wage. One-third of the pre- mium is paid by employees and two-thirds by employers. Self-employed workers pay the same total levy (13.5 percent) but only on 35 percent of their profits, subject to a legally defined minimum contribution (in 2004, this figure was CZK 905; the amount is adjusted periodically for infla- tion). The 80 percent of the self-employed who declare no profits pay only this minimum. The Ministry of Finance pays health insurance pre- miums on behalf of the nonactive population, including pensioners and people receiving social assistance, who together represent 56 percent of the total population. Premiums are computed by applying the same The Czech Republic 125 13.5 percent contribution rate to an amount established by a statutory order (CZK 3,458 in 2003). Voluntary health insurance is also available for those seeking coverage beyond what is provided under the mandatory scheme, although there is virtually no participation. As a result, voluntary insurance accounts for only 0.1 percent of total expenditures on health care. Institutional Structure and Coverage of Earnings-Related Scheme Under the supervision of the Ministry of Labor and Social Affairs, the Social Security Administration manages the first-pillar earnings-related pension scheme. The Social Security Administration collects contributions and pays benefits through a central office and 76 district offices located throughout the country. The scheme covers the entire economically active population, including the self-employed. Participation is voluntary for cer- tain categories of individuals, including people employed abroad. In 2006, 4.85 million people were enrolled in the scheme (roughly 66 percent of the total working-age population and 93 percent of the labor force). The same year, 2.68 million people were receiving pensions (roughly 26 percent of the population) (see Czech Social Security Administration 2006 and the World Bank's Statistical Information Management & Analysis database). Structure of Benefits The first-pillar earnings-related pension scheme provides old-age, disabil- ity, and survivorship pensions. The provisions governing each of these types of benefits are discussed as follows. Old-age benefits. To be eligible for an old-age pension, applicants must have worked for 25 or more years and have reached the minimum retire- ment age. Retirement with 15 years of service is allowed at age 65. Old-age pensions are the sum of a flat benefit and an earnings-related benefit based on an individual's earnings history. The earnings-related benefit accrues at a rate of 1.5 percent for each year of service applied to a pension assessment base (that is, the value of wages used in the computation of benefits), which provides higher levels of income replacement for lower levels of wages. Retirement ages have been increasing since 1996 at the rate of two months per year for men and four months per year for women. By 2030, the statu- tory retirement age will reach 65 for men, childless women, and women with one child; 64 for women with two children; 63 for women with three children; and 62 for women with more than three children. 126 Adequacy of Retirement Income after Pension Reforms Early retirement of up to three years is currently allowed for people with 25 or more years of service (by 2018, this will increase to 30 years), subject to the following reduction in benefits. For each 90 days of early retirement, the pension assessment base is reduced by 0.9 percent. The 2008 reform increases that reduction to an (almost) actuarially neutral 1.5 percent if retirement takes place 720 days before the statutory retire- ment age. All reductions are permanent (that is, they continue even after a pensioner reaches the normal retirement age). After 2020, it will be pos- sible to retire four years before statutory retirement age. After 2030, it will be possible to retire five years before statutory retirement age. People who work beyond the normal retirement age are awarded additional credit equal to 1.5 percent of their pension assessment base for each 90 days they defer their pension. Disability benefits. Disability benefits are awarded to individuals who suffer impairment that affects their ability to work. Full disability bene- fits are awarded to people who have lost 66 percent or more of their effective working capacity; partial disability benefits are awarded to peo- ple who have lost 33­66 percent of their working capacity. Benefits are subject to vesting requirements (table 4.5). The benefits paid to disabled people are the sum of a flat benefit (which is the same as that provided to old-age pensioners, regardless of whether a person is fully or partially Table 4.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in the Czech Republic under the First-Pillar Earnings-Related Scheme Partial Vesting period Contributions Eligibility Benefit ratea pension Under age 20: Less than 1 year No specific Full disability: Full disability: Flat Age 20­22: 1 year contribution at least 66 flat benefit benefit Age 22­24: 2 years rate for percent loss plus 1.5 plus 0.75 Age 24­26: 3 years disability of capacity percent percent Age 26­28: 4 years benefits, Partial per year per year Over age 28: 5 years no ceiling disability: Partial on wages at least 33 disability: subject to percent loss flat benefit contributions of capacity plus 0.75 percent per year Sources: European Commission 2005, 2007a; Czech Social Security Administration Web site (accessed March 2008). a. The 2008 reform introduced a new classification based on three degrees of invalidity, with reduced benefits for the first degree; details were unavailable in time to be included in this chapter. The Czech Republic 127 disabled) plus an earnings-related benefit. The earnings-related benefit is computed on the basis of a rate of accrual of 1.5 percent (for people who are fully disabled) or 0.75 percent (for people who are partially disabled). Regardless of the results of the benefit computation, disabled people are guaranteed a minimum pension equal to the minimum pen- sion paid to old-age pensioners. In addition, upon reaching the normal retirement age, pensioners are given the choice between receiving their disability pension or their old-age pension, whichever provides them with higher benefits. Survivor benefits. Survivor benefits are awarded to dependents if the deceased had been receiving (or had met the criteria to receive) an old- age or disability pension. Widows and widowers receive a flat benefit plus 50 percent of the earnings-related component of the deceased's pension (table 4.6). Widows over age 55 and widowers over age 58 are eligible to receive the pension for longer than one year; other widows and widowers receive the pension for only one year. Survivors who are disabled and survivors who are caring for a dependent disabled child or disabled parent are eligible for benefits irrespective of their age. Orphans receive a flat benefit plus 40 percent of the earnings-related component of the deceased's pension for each dependent child under age 26. For dis- abled survivors, the value of the pension is increased by 50 percent (for fully disabled survivors) or 20 percent (for partially disabled survivors). As a result of these comparatively generous provisions, total survivor benefits can actually exceed the pension to which the deceased was originally entitled. Assessment of the Performance of the Czech Pension System The World Bank has established four principles for evaluating public pen- sion systems, which together should guide the process of pension reform (see Holzmann and Hinz 2005). Broadly speaking, these principles include the adequacy and security of benefits, the affordability of contributions, the sustainability of the system over time, and the robustness of the sys- tem in the face of demographic changes and macroeconomic shocks. This chapter focuses primarily on the adequacy of benefits and financial sus- tainability of the first-pillar earnings-related pension scheme. The remain- ing principles are mentioned only briefly. Adequacy is analyzed through the lens of net replacement rates. Financial sustainability is evaluated using projections of pension expenditures and revenues. 128 Table 4.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in the Czech Republic under the First-Pillar Earnings-Related Scheme Spouse Orphan age Orphan Eligibility replacement rate Benefit duration Remarriage test limit replacement rate Total family benefit Eligibility of Flat benefit plus 50 For life, if spouse is 70 Pension ceases if 26 Flat benefit plus 40 No maximum deceased for percent of percent disabled, survivor remarries percent of deceased's old-age or deceased's is taking care of a pension disability pension child or a pension dependent parent, or is age 55 (women) or age 58 (men); otherwise, one year Sources: European Commission 2005, 2007a; Czech Social Security Administration Web site (accessed March 2008). The Czech Republic 129 Benefit Adequacy Replacement rates are a useful yardstick for measuring the adequacy of pension benefits because they express benefits relative to preretirement earnings, thereby demonstrating the degree to which income is replaced when workers retire. Two variants are commonly used. Gross replace- ment rates compute income replacement as the ratio of benefits paid to pretax preretirement earnings. Net replacement rates compute income replacement as the ratio of benefits received (that is, after the payment of taxes and other levies, including contributions for social insurance) to posttax preretirement earnings. In general, net replacement rates are a more useful measure of benefit adequacy because they capture the degree to which actual take-home pay is replaced when workers retire. Benefit adequacy is determined not only by the level of income replacement at retirement. For a full assessment of benefit adequacy, it is also important to assess how postretirement indexation rules will affect replacement rates during retirement. Pension benefits in retire- ment are expected to be indexed to inflation so that their real value is maintained. In a growing economy with increasing real wages, mere price indexation of pensions, however, leads to a deterioration of the relative consumption position of the retirees. Individuals with other- wise identical work histories will receive different pensions depending on when they retire. For this reason, some countries have introduced mixed indexation of pensions with varying weights of inflation and wage growth in the indexation formula. For an evaluation of the effect of indexation on replacement rates in the Czech Republic, the replacement rates are normalized to 100 and the assumptions for calculating the replacement rates are maintained (that is, inflation is 2.5 percent a year and real wage growth is 2 percent a year). The change in the replacement rate is measured in comparison to full wage indexation or compared to an active worker. The results of this analysis indicate that the relative income position of a retiree would dete- riorate by 12 percent after 10 years in retirement and by 36 percent after 35 years in retirement. Because benefit indexation in the Czech Republic includes an adjustment amounting to one-third of wage growth over inflation, this deterioration is lower than it is in countries that fully price index benefits. The evaluation of income replacement that follows con- siders replacement rates only at retirement and does not take into account the impact of indexation policies on replacement rates during retirement. Replacement rates are a function of the formula governing pension benefits; an individual's contribution history; and, in the case of the net 130 Adequacy of Retirement Income after Pension Reforms replacement rates, the rules of income tax, social security contributions, and other relevant levies. The benefit formula establishes the degree to which the system redistributes income across individuals of different lev- els of preretirement earnings. Progressive systems provide higher levels of income replacement to people with lower levels of preretirement income. In general, the degree to which a system is redistributive depends on the existence (and value) of flat transfers and minimum pension guarantees, the degree to which benefits are earnings related, and the existence of ceilings on earnings subject to contributions. An individual's contribution history can be characterized by his or her age of entry into the labor force, contribution density, and decisions regarding the timing of retirement. To some degree, these three factors are influenced by the incentives embod- ied in the pension system. The tax and contribution system influences net replacement rates through the typical progressiveness of the income tax formula, which taxes higher income during a worker's active life more so than it does lower pension benefits in retirement. In addition--and, even more important, for low to middle-income groups--there are social secu- rity levies (for pensions; unemployment; health care; and, at times, hous- ing and family benefits), which are typically reduced (for example, health care) or eliminated altogether in retirement. The adequacy of income replacement provided by the first-pillar earnings-related pension scheme in the Czech Republic cannot be evalu- ated without first establishing benchmarks. Unfortunately, there is no consensus on what constitutes adequacy.According to one widely respected definition, pensions are adequate when they are sufficient to prevent poverty among the elderly and provide the vast majority of the popula- tion with a reliable mechanism for smoothing income over their lifetime. Even with the benefit of a definition, however, establishing benchmarks is problematic because attitudes vary from one country to another as a function of social and cultural perceptions. Moreover, benchmarks ignore the existence of other factors that affect the welfare of the elderly--and that vary from country to country--including the existence and generos- ity of health insurance and long-term care, the cost of housing, the struc- ture of traditional living arrangements, the presence of informal intrafamily or intergenerational sources of financial and nonfinancial sup- port, and the availability and security of other mechanisms for people to save for their own retirement. One reputable nine-country study (OECD 2001) observed that liv- ing standards are roughly comparable for people 10 years older than the normal retirement age and people 15 years younger than the normal retirement age when retirees have disposable income equal to roughly The Czech Republic 131 80 percent of the disposable income of working-age people. In part, this is attributable to the fact that retirees have no work-related expenses (that is, they do not have to commute to and from a place of employ- ment or buy special clothing or uniforms, and so forth). This finding, however, does not imply that mandatory first-pillar pension schemes should actually target an 80 percent net replacement rate. To the con- trary, in middle- and high-income countries, one can reasonably expect individuals to save for their own retirement--and the empirical evidence suggests that, in practice, this is actually happening.8 There is also some evidence to suggest that the ratio between preretirement and postretire- ment income is somewhat independent of the income replacement man- date of the public pension system. Put simply, individuals tend to save more in countries with more modest mandates (and vice versa). Because the Czech Republic has access to relatively well-developed financial markets, it would seem reasonable to expect middle- and higher- income workers to save enough to finance at least 25 percent, if not closer to 50 percent, of this 80 percent income replacement target. Given this, three benchmarks are provided: a 40 percent net replacement rate (which implies that individuals would be expected to save enough to finance half of the total income replacement target); a 60 percent net replacement rate (which implies that individuals would be expected to finance a quar- ter of the target); and an 80 percent net replacement rate (which implies that individuals, most of whom would be low-income earners, would not be expected to contribute anything toward the target).9 In the following analysis, these three benchmarks are used to evaluate the adequacy of benefits in the Czech Republic compared with the average net replace- ment rate observed in 53 countries around the world, the average net replacement rate observed for selected countries in Europe and Central Asia, and the poverty line in the Czech Republic. To estimate gross and net replacement rates, we consider two critical dimensions--earnings levels and contribution periods--with the help of the Analysis of Pension Entitlements across Countries (APEX) model.10 This model generates estimates for replacement rates under steady-state assump- tions (that is, as if the rules of the reformed pension scheme had been in place over the entire active life of the individual). Because life expectancies at retirement are projected to increase over time--which will affect the ben- efits paid by defined-contribution pension schemes--a reference year must be chosen. For this study, 2040 is used, because it provides a sufficiently long contribution period over which to approximate steady-state conditions. The first critical task is to investigate levels of income replacement across a relevant spectrum of income. Income is represented as a percentage 132 Adequacy of Retirement Income after Pension Reforms (50­200 percent) of average earnings. The second task is to investigate the impact on income replacement of differences in the duration, timing, and density of an individual's contribution history (density refers to the percent- age of time an individual actually contributes over a given period). To facil- itate the presentation of these multidimensional results, replacement rates are computed as a function of the age an individual exits the labor market. They are presented separately for full-career and partial-career workers. Replacement rates for full-career workers. For the purpose of this analy- sis, a full career is defined as continuous employment from age 20 to the current normal retirement age of 63. Gross replacement rates clearly show why the Czech Republic's first-pillar earnings-related pension scheme has been described as highly progressive (figure 4.1). The level of gross income replacement provided to someone earning half the average wage is more than twice that provided to someone earning twice the average wage. This is, of course, caused by the presence of the redistributive earnings-related benefit, as well as the flat benefit, which contributes less to total benefits as income rises. The situation does not change significantly when taxes are taken into consideration. As a result of the impact of taxes and contributions, net replacement rates fall only slightly more than do gross replacement rates when expressed relative to preretirement income (figure 4.2). Figure 4.1 Sources of Gross Replacement Rates in the Czech Republic, by Income Level 80 70 60 (percent) 50 rate 40 30 replacemnt 20 10 gross 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage earnings-related benefit flat benefit Source: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. The Czech Republic 133 Figure 4.2 Sources of Net Replacement Rates in the Czech Republic, by Income Level 100 90 (percent) 80 rate 70 60 50 40 replacement 30 of 20 10 0 proportion 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage earnings-related benefit flat benefit taxes Source: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. Net replacement rates for full-career workers in the Czech Republic can generally be considered adequate (figure 4.3).11 Replacement rates are substantially higher than the highest of the three benchmarks for peo- ple with low preretirement income, slightly higher than the middle benchmark for people with roughly average preretirement income, and roughly equal to or higher than the lowest benchmark for people with high preretirement income. This finding suggests that the system is effec- tively smoothing consumption from work to retirement, more so for lower-income workers, who have less capacity to save on their own. Given that benefits for even the lowest full-career workers greatly exceed the poverty line, the objective of poverty alleviation is being met.12 The pension system in the Czech Republic is more redistributive than the average pension system in the region (or the world), as demonstrated by the fact that it provides higher levels of income replacement for lower- income workers and lower levels of income replacement for higher- income workers, with respect to regional and world averages. Replacement rates for partial-career workers. Not everyone works from age 20 to the statutory retirement age; many individuals enter and exit the labor force (often at different ages and for different periods of time). To examine the adequacy of benefits for partial-career workers, three stylized cases are considered. These include career type A (someone entering the 134 Adequacy of Retirement Income after Pension Reforms Figure 4.3 Net Replacement Rates for Male Full-Career Workers in the Czech Republic, Europe and Central Asia, and the World, by Income Level 120 high benchmark 100 (percent) 80 rates 60 middle benchmark 40 low benchmark replacement net 20 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage world average Europe and Central Asia average Czech Rep. average poverty line (percentage of average income) Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as an approximation of steady-state conditions. labor force at age 25 who works continuously for a period of years before leaving the workforce at some point between the ages of 50 and 70 and then claims a benefit); career type B (identical to career type A, except that the worker enters the workforce at age 30 and leaves no earlier than age 55); and career type C (identical to career type A, except that the individ- ual contributes in only three years out of four while in the labor force). In cases in which the withdrawal from the formal labor market occurs before the statutory retirement age, the pension is claimed (and the replacement rate calculated) only at the later age. For withdrawals after the statutory retirement age, of course, the ages coincide. Net replacement rates for low-income partial-career workers are examined first (figure 4.4). Because a minimum of 25 years of contribu- tions are required to become eligible for old-age pension benefits, the age at which these workers can permanently exit the workforce and still claim a benefit can vary (this explains why the lines for each of the three career types start at different ages). The acceleration in the increase in The Czech Republic 135 Figure 4.4 Net Replacement Rates for Male Low-Income Partial-Career Workers in the Czech Republic, by Career Type and Exit Age 150 world average 140 130 high benchmark 120 110 100 (percent) 90 rate 80 Europe and Central Asia average 70 middle benchmark 60 50 low benchmark 40 replacement poverty line (percentage 30 of average income) 20 net 10 0 45 50 55 60 65 70 exit age from labor market career type A: career type B: career type C: entry age--25, entry age--30, entry age--25, contribution contribution contribution density--100% density--100% density--75% Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as an approximation of steady-state conditions. See text for descriptions of career types. replacement rates with years of contributions after age 60 reflects the fact that each extra year's contributions increase benefits both directly and, because retirement is delayed, by a lower actuarial reduction or an actuarial increment. Two conclusions can be drawn from figure 4.4. First, entering the workforce later in life is costly for low-income workers (someone enter- ing the workforce at age 30 receives a net replacement rate that is some 15 percentage points lower than someone entering the workforce at age 25). Second, working intermittently is costly (someone entering the work- force at the same age but who contributes only three years out of four will receive a net replacement rate that is 9 percentage points lower than someone who contributes continuously). Figure 4.4 also suggests that some low-income partial-career workers--particularly those who exit the labor market long before reaching the statutory retirement age--may not receive acceptable levels of income replacement (their net replacement rates are lower than the lowest of the benchmarks). Some low-income workers (career types A and B) who work until the statutory retirement 136 Adequacy of Retirement Income after Pension Reforms age (which is likely to include people with limited ability to save on their own, for whom the mandatory system is likely to be their main source of retirement income) receive a net replacement rate of more than 80 percent, while others (career type C) receive a replacement rate of less than 80 percent. The conclusions about net replacement rates for middle-income partial- career workers are similar, except that the differences in net replacement rates across the three types of workers are smaller, because the contribution of the flat benefit declines as preretirement income increases (figure 4.5). Entering the workforce later in life is still costly for middle-income work- ers but less so than for low-income workers (the difference between career types A and B is only 6 percentage points). Working intermittently is also costly for middle-income workers but less so than for low-income workers (the difference is only 4­8 percentage points). Some middle- income partial-career workers--including those who exit the labor market long before reaching the statutory retirement age--will not receive accept- able levels of income replacement, because their net replacement rates may be beneath the poverty line. To earn income replacement that is Figure 4.5 Net Replacement Rates for Male Middle-Income Partial-Career Workers in the Czech Republic, by Career Type and Exit Age 100 high benchmark 90 world and Europe and Central Asia average 80 70 (percent) middle benchmark 60 rate 50 low benchmark 40 poverty line (percentage of 30 average income) replacement 20 net 10 0 45 50 55 60 65 70 exit age from labor market career type A: career type B: career type C: entry age--25, entry age--30, entry age--25, contribution contribution contribution density--100% density--100% density--75% Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. The Czech Republic 137 higher than the lowest benchmark, for example, middle-income workers must work until age 63, depending on their age of entry into the work- force and their contribution density. To earn income replacement of 60 percent of preretirement income, they must work at least until (and, in some cases, one to two years beyond) the normal retirement age. To achieve income replacement of 80 percent, they must work until at least age 68--five years longer than low-income workers need to work to earn the same level of income replacement. The results for high-income partial-career workers are similar (figure 4.6). Workers who leave the workforce at earlier ages may receive replacement rates below the poverty line. The Czech pension system still provides high-income partial-career workers with an acceptable level of income replacement, albeit at lower levels than that provided to either middle- or low-income workers, because the benefit formula is progressive. To earn income replacement that exceeds the lowest benchmark, for example, high-income workers must work until at least age 66 (depending on their age of entry into the workforce and their contribution density). Even if high-income workers continue working until age 70, however, they Figure 4.6 Net Replacement Rates for Male High-Income Partial-Career Workers in the Czech Republic, by Career Type and Exit Age 90 high benchmark 80 world and Europe and Central Asia average 70 middle benchmark 60 (percent) 50 rate 40 low benchmark 30 20 replacement poverty line (percentage net of average income) 10 0 45 50 55 60 65 70 exit age from labor market career type A: career type B: career type C: entry age--25, entry age--30, entry age--25, contribution contribution contribution density--100% density--100% density--75% Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as an approximation of steady-state conditions. 138 Adequacy of Retirement Income after Pension Reforms will not earn income replacement of 60 percent of their preretirement income, which implies that these workers must save outside of the pen- sion system in order to preserve their standard of living into retirement. The key conclusion to be drawn from this analysis is that the Czech pension system generally provides benefits that are adequate to allevi- ate poverty for all workers but that partial-career workers who leave the workforce before reaching the statutory retirement age and those with intermittent work histories may receive replacement rates that leave them below the poverty line. The Czech pension system provides full-career workers with a reliable means of preserving their standard of living into retirement (that is, lifetime consumption smoothing), but some partial-career workers, particularly those with higher incomes, will need to save outside the pension system in order to accomplish the same objective. Fiscal Sustainability The sustainability of a pay-as-you-go first-pillar pension scheme is best evaluated in actuarial terms by estimating the scheme's actuarial deficit as the difference between its assets and liabilities. If an actuarial deficit exists and is large, the scheme is financially unsustainable and needs pol- icy actions that increase assets, reduce its liabilities, or both. A good proxy for the actuarial deficit is the difference between the present value of the scheme's expected future revenues (that is, contributions and other income) and expected future expenditures (that is, benefit payments, administrative costs, and other expenses) over an extended projection period. The difference between these two values represents an unfunded liability (sometimes referred to as a financing gap) on the public-sector balance sheet. Because this study is concerned also with the time path of revenues and expenditures (and the resulting balance across the projec- tion period ending in 2050), this more pragmatic approach has been taken, and projections of expenditures, revenues, and deficits are pre- sented on the basis of available postreform fiscal projections. The reforms of 1995 and 2003 and those introduced in 2008 are expected to improve the fiscal position of the first-pillar earnings-related pension scheme. They did not, however, ensure the long-term sustainabil- ity of the scheme, which is projected to generate a slight surplus until 2032, with revenues and expenditures remaining roughly constant at about 8 percent of GDP (figure 4.7). Starting in 2033, the gap between revenues and expenditures is projected to gradually widen, leading to a deficit equivalent to 2.5 percent of GDP by 2050.13 The Czech Republic 139 Figure 4.7 Projected Fiscal Balance of the Public Pension System in the Czech Republic after Reform, 2009­50 12 10 8 GDP 6 of 4 2 percentage 0 ­2 ­4 2009 2014 2019 2024 2029 2034 2039 2044 2050 year expenditures revenues deficit Source: Czech Ministry of Labor and Social Affairs 2008. Figure 4.8 Projected Old-Age and System Dependency Ratios in the Czech Republic, 2005­50 100 90 80 70 60 50 (percent) 40 ratio 30 20 10 0 2005 2010 2020 2030 2040 2050 year system dependency ratio old-age dependency ratio Sources: European Commission 2005; Reiterer 2008. The aging of the population is largely to blame for these projections. As in other countries in the region--many of which are also facing rela- tively low rates of fertility in combination with increased life expectancy--the population in the Czech Republic is aging rapidly. Demographic projections indicate that the old-age dependency ratio will rise from 21.7 percent in 2005 to 60.1 percent by 2050 (figure 4.8).14 140 Adequacy of Retirement Income after Pension Reforms The aging of the population is also driving an increase in the pension scheme's system dependency ratio.15 Although the system dependency ratio--which is now 2.5 times higher than the old-age dependency ratio--is projected to decline relative to the old-age dependency ratio (as a result of the reforms already undertaken), the aging of the population will eventually drive the pension scheme into insolvency. What options exist for restoring the system to fiscal balance? Unfortunately, for policy makers, the options are limited. Revenues can be increased by raising the contribution rate. Alternatively (or in addition to--the options are not mutually exclusive), expenditures can be reduced by cutting benefits, increasing the minimum number of years required to become eligible for benefits, or delaying the payment of benefits by rais- ing the retirement age further. Because raising the contribution rate could threaten competitiveness and will likely strengthen incentives for tax eva- sion, it is typically not embraced by policy makers. This leaves policy makers with limited options: cutting benefits, tight- ening eligibility conditions, or raising the retirement age. It also raises the question of whether restoring sustainability will exact a cost in terms of the adequacy of benefits provided to future beneficiaries. A rough analysis indicates that restoring sustainability to the first-pillar earnings- related pension scheme will require either increasing the retirement age to at least 66 by 2050 for both men and women16 or cutting bene- fits by an average of 23 percent. A 23 percent reduction in benefits results in a drop in net income replacement of 6­15 percentage points by 2050 (figure 4.9). A comparison of these new (lower) net replacement rates against the three benchmarks suggests that a 23 percent reduction in benefits would not cause income replacement to fall below the poverty line for full-career workers. This indicates that a sustainable first-pillar pension scheme in the Czech Republic will still achieve its poverty alle- viation objective. However, the same reduction in benefits would frus- trate the objective of smoothing lifetime consumption for full-career workers, because levels of income replacement would be lower than the 80 percent benchmark for low-income workers, lower than the 60 percent benchmark for middle-income workers, and lower than the 40 percent benchmark for high-income workers. This last observation is subject to three caveats. First, these findings may be unduly pessimistic if coverage is extended (because high rates of coverage will increase the resources available in the medium term for funding benefits, albeit at the cost of higher benefits in the long term) or if retirement ages are also increased (because delaying retirement can The Czech Republic 141 Figure 4.9 Net Replacement Rates for Male Full-Career Workers in the Czech Republic before and after Benefit Adjustment, by Income Level 110 100 90 80 high benchmark (percent) 70 rate 60 middle benchmark 50 40 low benchmark 30 replacement 20 net 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage net replacement rate adjusted net replacement rate poverty line (percentage of average income) Source: Authors' calculations based on World Bank 2007a and the APEX model. reduce the rates of return paid on contributions, depending on how a pension scheme is structured). Second, high-income workers always have the option of saving outside of the first-pillar pension scheme. To increase income replacement by 1 percentage point, for example, a full- career worker would need to save only about 0.5 percent of his or her earnings from age 40 to the current age of retirement.17 Third, this analy- sis considers only full-career workers, while the average worker now con- tributes for only about 30 years, substantially less than the 40 years expected over a full career. Contributing to the pension scheme for only 30 years, for example, reduces net income replacement by 18 percentage points (for low-income workers) and 8 percentage points (for high- income workers). Conclusions The Czech Republic is one of only a few countries in the region that did not look beyond its first-pillar pay-as-you-go pension scheme when it reformed its pension system in 1995, 2003, and 2008. The reforms enacted to date have been parametric, except for the introduction, in 1994, of a budget-subsidized voluntary third-pillar scheme. These parametric 142 Adequacy of Retirement Income after Pension Reforms reforms improved the fiscal condition of the scheme, but they did not make the scheme sustainable over the long term, largely because they failed to go far enough to counteract the impact of the aging of the Czech population. The first-pillar pension scheme is highly progressive, in terms of both its levels of income replacement and the rates of return it provides on contributions. The degree to which the scheme provides more generous income replacement and higher returns to workers with lower levels of preretirement income is attributable to a redistributive earnings-related benefit as well as to the existence of a flat benefit, which contributes less to total benefits as incomes rise. The resulting gross and net replacement rates for full-career workers at average earnings levels are fully in line with regional and international standards. For workers at lower and higher earnings levels, however, replacement rates deviate substantially from these standards as a result of the redistributive benefit formula (see chapter 1). First-pillar pension benefits are adequate to alleviate poverty for all workers: both full-career and partial-career workers are receiving levels of income replacement well above the poverty line. The scheme is also successfully smoothing lifetime consumption for all full-career workers. Many partial-career workers, however, especially those earning more than the average wage, will have to work beyond even the higher future retirement age of 65 (applicable to men, childless women, and women with one child) or save outside the pension scheme to preserve their standard of living into retirement. Increasing the retirement age in step with increases in life expectancy at retirement is a natural choice for both individuals and pol- icy makers, but it requires cross-sectoral policy reforms to enable eld- erly workers to continue to participate in the labor market.18 In the future, Czech policy makers will have to either allocate funds from the budget to finance deficits in the first-pillar pension scheme or enact further reforms to restore fiscal balance to the scheme. Restoring fiscal balance to the scheme while successfully smoothing lifetime con- sumption (particularly for middle- and high-income workers) will be dif- ficult to accomplish on the basis of benefit cuts alone. Moreover, unless increases in the statutory retirement age are accompanied by changes in the benefit formula (such as a reduction in the accrual rate, as enacted by the 2003 and 2008 reforms), such increases will only lengthen the aver- age pension contribution period, thereby leading to higher replacement rates and increased expenditures. A complementary option to reducing benefits would be to change to pure price indexation. The Czech Republic 143 The Czech government recognizes that the population is aging. It has been subsidizing the voluntary pension scheme in order to supplement the retirement income provided by the first-pillar scheme. These subsi- dies explain the comparatively broad coverage enjoyed by the voluntary scheme. However, the low levels of average contributions being made suggest that the scheme will not provide a meaningful benefit (or income replacement) for most participants. Of equal concern is the fact that participation in the voluntary scheme is much lower for younger work- ers, who are more vulnerable to reforms in the first-pillar scheme that reduce their benefits (but who stand to gain the most by virtue of having longer investment horizons). Clearly, there is scope for the expansion and growth of the voluntary pension scheme in the provision of retirement income. This aspect of the overall pension system in the Czech Republic requires--and deserves--further investigation. Notes 1. For an overview of the main measures adopted by the Czech Republic since 1990 in the area of social insurance, see Czech Republic Ministry of Labor and Social Affairs (2006). 2. The benefit increases by 2,880 koruny (CZK) for the first additional person in a household and by CZK 2,600 for other additional persons who are not depend- ent children. It rises by CZK 1,600 for each dependent child under age 6, CZK 1,960 for each dependent child age 6­15, and CZK 2,250 for each dependent child age 15­26 (European Commission 2007a;World Bank 2007a). 3. In 2006 the flat benefit was CZK 1,470, or 7.3 percent of the average wage. The Czech pension system also guarantees a minimum pension equal to the sum of the flat benefit and CZK 770 in 2006. 4. The first CZK 9,100 per month is replaced at 100 percent; earnings of CZK 9,100­CZK 21,800 are replaced at 30 percent; earnings above CZK 21,800 are replaced at 10 percent. 5. This change is being implemented at the rate of one additional year of wages per year (currently, 18 years of wages are used). Historical wages are adjusted (that is, they are valorized by average wage growth). 6. Lump-sum payments accounted for 85 percent of total pension payments in 2004 (Czech National Bank 2004). 7. In 2004, the number of participants was 2,963,730 (Czech National Bank 2004). 8. In Chile, for instance, 70 percent of retirees from the mandatory public pension system own their home, which, of course, is a form of savings (see 144 Adequacy of Retirement Income after Pension Reforms Valdés-Prieto 2008). In the Czech Republic, the rate of home ownership among retirees is of similar magnitude. 9. These benchmarks approximate the standards developed by the International Labour Organization (ILO) (1952) and the Council of Europe (1990). ILO Convention 102 of 1952 sets a minimum benefit equal to 40 percent of the reference wage for married men of pensionable age. This amount was raised to 45 percent in 1968. The European Code of Security of 1990 sets a min- imum standard for members of the Council of Europe equal to 65 percent for married people of a specific age. 10. The APEX model was developed by Axia Economics, with funding from the OECD and the World Bank. The model codes detailed eligibility and benefit rules for first- and second-pillar schemes based on available pub- lic information that has been verified by country contacts. Because the details of the rules sometimes change on short notice (and limited public disclosure), the calculations presented here should be considered as best approximations only. 11. These are simulated replacement rates for unmarried men based on a hypo- thetical career path and the assumptions that real wage growth is 2 percent, inflation is 2.5 percent, and everyone retires at the statutory retirement age. 12. As a proxy for the poverty line, this study uses 35 percent of the average net wage, because this percentage very broadly approximates a US$2.25-a-day poverty line converted into national currency, adjusted for purchasing power parity, expressed relative to the national average net wage, and averaged across the nine study countries. Such an approach enables valid comparisons to be made across the sample (see chapter 1). 13. These projections by the Ministry of Labor and Social Affairs are based on a somewhat different methodology and assumptions than were those from the Ministry of Finance, which served as an input to the EU Economic Policy Committee--Ageing Working Group (European Commission 2007b). As a result of these differences, these projections estimate the deficits to be lower by about 1 percent of GDP. 14. The old-age dependency ratio is the population age 65 and higher divided by the population age 20­64. 15. The system dependency ratio is the number of people receiving a pension divided by the number of people contributing to the pension scheme. 16. This estimate is based purely on demographic projections. It assumes that everyone age 65 and older receives a pension, everyone age 20­65 contributes, and all pensioners receive the replacement rate awarded to the average worker. 17. This estimate is based on the assumption that real wage growth is 2 percent, the net real rate of return on invested assets is 3.5 percent, and benefits (from both the unfunded and the funded pillars) are price indexed. Country- specific mortality rates are used. The Czech Republic 145 18. See Holzmann, MacKellar, and Repansek (2009) for a conference volume that addresses theses issues for the countries in southeastern Europe. Bibliography Council of Europe. 1990. European Code of Social Security (Revised). Rome. Czech Republic Ministry of Labor and Social Affairs. 2006. Actuarial Report on Social Insurance. Prague. Czech National Bank. 2004. Annual Report. Office of the State Supervision in Insurance and Pension Funds, Prague. Czech Social Security Administration. n.d. http://www.mpsv.cz/cs/. ------. 2006. Annual Report. Prague. European Commission. 2005. Employment and Social Affairs. National Strategy Report on Adequate and Sustainable Pensions. Brussels. ------. 2007a. MISSOC (Mutual Information System and Social Protection) database. http://ec.europa.eu/employment_social/. ------. 2007b. Pension Schemes and Projection Models in EU-25 Member Countries. European Economy Occasional Paper 37, Economic Policy Committee and Directorate General for Economic and Financial Affairs, Brussels. GVG (Gesellschaft für Versicherungswissenschaft und ­gestaltung) and European Commission. 2003. Study on the Social Protection Systems in the 13 Applicant Countries: Czech Republic Country Report. Cologne. Hemmings, P., and E.Whitehouse. 2006. Assessing the 2005 Czech Proposal for Pension Reform. Paris: Organisation for Economic Co-operation and Development. Holzmann, R., and R. Hinz. 2005. Old Age Income Support in the 21st Century. Washington, DC: World Bank. Holzmann, R., L. MacKellar, and J. Repansek, eds. 2009. Pension Reform in South- eastern Europe: Linking to Labor and Financial Market Reforms. Washington, DC: World Bank. Iglesias, A. 2003. "Strengthening the Private Voluntary Pension Scheme in the Czech Republic." http://www.zavedenieura.cz/cps/rde/xbcr/mfcr/5-SB.pdf. ILO (International Labour Organization). 1952. ILO Convention 102. Geneva: ILO. ------. 1967. ILO Convention 128. Geneva: ILO. Lasagabaster E., R. Rocha, and P. Wiese. 2002. "Czech Pension System: Challenges and Reform Options." Social Protection Discussion Paper 0217, World Bank, Washington, DC. Laursen, T. 2000. "Pension System Viability and Reform Alternatives in the Czech Republic." Working Paper, International Monetary Fund, Washington, DC. Lindbeck, A., and M. Persson. 2003. "The Gains from Pension Reform." Journal of Economic Literature 41 (March): 74­112. 146 Adequacy of Retirement Income after Pension Reforms Natali, David. 2004. "The Reformed Pension System." Observatoire Social Européen. OECD (Organisation for Economic Co-operation and Development). 2001. Ageing and Income: Financial Resources and Retirement in 9 OECD Countries. Paris: OECD. Reiterer, A. 2008. "Population Development and Age Structure in Southeastern Europe until 2050." World Bank, Washington, DC. Rokosova, M., and B. Schreyogg. 2005. Healthcare Systems in Transition: Czech Republic. World Health Organization, Regional Office for Europe, Copenhagen. U.S. Social Security Administration. 2006 Social Security Systems throughout the World: Europe. Washington, DC: Social Security Administration. Valdés-Prieto, S. 2008. Designs for the First-Pillar Pensions and the 2008 Chilean Reform. http://editorialexpress.com/cgi-bin/conference/download.cgi?db_ name=SECHI2008&paper_id=130. World Bank. 2007a. Pensions Panorama. Washington, DC: World Bank. ------. 2007b. Social Assistance in Central Europe and the Baltic States. Washington, DC: World Bank. WHO (World Health Organization). 2008. Database. http://www.who.int/ research/en/. C H A P T E R 5 Hungary Hungary inherited a socialist-era defined-benefit pension system financed on a pay-as-you-go basis (meaning that contributions from current work- ers are used to pay benefits to current beneficiaries). Within a few years of the country's transition to a market economy, the fiscal balance of the pension system began to deteriorate. The system broke even in 1992 but generated deficits thereafter that were projected to continue to grow. Despite the system's high contribution rate, pension revenues increas- ingly fell short of expenditures, as a result of declining employment and increasing informality in the labor market. At the same time, disability and old-age pension provisions were used to cushion rising unemployment. Long-term projections showed that the fiscal balance of the system would worsen over time, with deficits reaching 6 percent of gross domestic prod- uct (GDP) by 2050, driven by the aging of the Hungarian population. Recognizing that low retirement ages, lax benefit eligibility condi- tions, and poor levels of compliance (resulting from high payroll taxes) were imposing short-term pressure on the fiscal balance of the pension system--and aware of the potentially devastating impact of the aging of the population over the long term--Hungarian policy makers launched a comprehensive package of reforms to the pension system in 1997­98. These reforms moved the system from a monopillar design 147 148 Adequacy of Retirement Income after Pension Reforms toward a multipillar design that included a mandatory fully funded defined-contribution scheme and introduced parametric changes to the existing pay-as-you-go defined-benefit scheme. Projections at the time of reform suggested that these reforms would improve the long-term fiscal balance of the pension system, with deficits reaching only 3 percent of GDP by 2050 rather than the 6 percent previously projected. Improving the long-term finances of the pension system further will require increas- ing employment across all ages, making benefits less generous, or raising retirement ages (or some combination thereof ). Against this backdrop, this chapter evaluates the Hungarian pension system, focusing on fiscal sustainability and benefit adequacy. Adequacy is evaluated through the lens of statutory net replacement rates for differ- ent retirement ages, patterns of contributions, and income levels relative to international benchmarks. This chapter is organized as follows. The next section discusses the motivation for the reforms. The following section describes the key char- acteristics of the reformed pension system. The third section assesses the adequacy of pension benefits and the fiscal sustainability of the system. The last section draws conclusions. Motivation for Reform Like other transition economies in the region, Hungary inherited a pen- sion system that suffered from a number of serious design flaws, includ- ing low retirement ages, liberal eligibility conditions, and high payroll taxes. By 1993, the combination of these factors resulted in a system deficit equivalent to 0.3 percent of GDP--a deficit that had to be funded using transfers from the central budget (table 5.1). The strongest driver behind the worsening fiscal condition of the pension system was a sharp drop in contribution revenue (from 11 percent of GDP in 1991 to 8.5 percent by 1996) as a result of (a) declining employment in the formal economy;1 (b) restrictive wage policies introduced by the government in 1995 as part of the stabilization program that led to a fall in real wages (and con- tributions to the pension scheme) of more than 10 percent that year (Palacios and Rocha 1998); (c) higher student enrollment ratios; (d) high rates of early retirement; and (e) a decline in the number of working pensioners. Average pensions and (consequently) expenditures also fell during this period, albeit not by as much as revenues. Between 1994 and 1996, the pension system experienced deficits of 0.4­1.7 percent of GDP, breaking even again 1997. Hungary 149 Table 5.1 Fiscal Balance of Hungary's Pension System before Reform, 1991­96 (percentage of GDP) Year Revenues Expenditures Balance 1991 11.0 10.5 0.5 1992 10.4 10.4 0.0 1993 10.1 10.4 ­0.3 1994 9.7 11.4 ­1.7 1995 8.9 10.5 ­1.6 1996 8.7 9.1 ­0.4a Source: Palacios and Rocha 1998. Note: Table includes revenues and expenditures for the entire system, including old-age, disability, and survivor pensions, as well as short-term benefits. Computing the balance of the system was confounded by the fact that health and pension benefits were funded by a common payroll tax. Deficits shown assume a notional contribution rate of 33.6 percent, sufficient to cover all expenditures in 1992 but insufficient to cover subsequent expenditure. a. Ministry of Finance data. Projections from 1998 indicate that before the reforms of the mid- to late 1990s, the pension system would have generated increasingly larger deficits, eventually reaching 6 percent of GDP by 2050--a burden that would be unaffordable and difficult to bear given competing demands on the government's resources (figure 5.1). Deficits were projected largely because of aging of the population. As in most other countries in the region, the population in Hungary is projected to age, albeit somewhat less rapidly.2 The old-age dependency ratio (the population age 65 and older divided by the population age 20­64) is projected to increase from 25 per- cent in 2005 to 53.3 percent by 2050.3 The aging of the population is expected to significantly increase the total number of beneficiaries relative to contributors. This is captured by the system dependency ratio (the num- ber of people receiving a pension divided by the number of people con- tributing to the pension scheme), which was projected to increase from 51 percent in 2005 to 123 percent by 2050 as the number of beneficiaries rises far more rapidly than the number of contributors (figure 5.2).4 Growing pension deficits in the pay-as-you-go defined-benefit pension scheme in the mid-1990s prompted policy makers to introduce a multipil- lar pension system, the primary objective of which was to improve long- term sustainability. The reforms--initiated in 1997 and implemented in 1998--moved the system from a monopillar design toward a multipillar design that included a mandatory, fully funded, defined-contribution scheme and introduced parametric changes to the existing pay-as-you-go defined-benefit scheme. These reforms were considered radical, because Hungary was the first country in the region to attempt to finance a portion of pension benefits on the basis of funded defined-contribution accounts. 150 Adequacy of Retirement Income after Pension Reforms Figure 5.1 Projected Fiscal Balance of Hungary's Public Pension System before Reform, 2000­50 0 ­1 ­2 deficit GDP of ­3 ­4 percentage­5 ­6 ­7 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 year Source: Palacios and Rocha 1998. Note: The figure includes revenues and expenditures for old-age pensions only. Figure 5.2 Projected Old-Age and System Dependency Ratios in Hungary, 2005­50 140 120 100 80 percent 60 40 20 0 2005 2010 2020 2030 2040 2050 year system dependency ratio old-age dependency ratio Sources: Palacios and Rocha 1998 (system dependency ratio); Reiterer 2008 (old-age dependency ratio). Characteristics of Hungary's Pension System This section describes the main characteristics of Hungary's pension sys- tem. These characteristics include the design of the individual pillars of social insurance; the rules governing pension system taxation, institutional Hungary 151 structure, and coverage; and the provisions governing old-age, disability, and survivorship pensions. The design of the pension system is compared using a conceptual framework developed by the World Bank, which generally recommends including a funded component if conditions are appropriate but increasingly recognizes that a range of choices is available to policy makers to provide effective old-age protection in a manner that is fiscally responsible (see Holzmann and Hinz 2005). In general, the World Bank supports pension systems composed of some combination of five basic pillars: · a noncontributory (or zero) pillar (in the form of a demogrant, social pension, or social assistance benefit) intended to provide a minimal level of income protection; · a first-pillar contributory system linked to earnings, which seeks to replace a portion of preretirement income; · a mandatory second pillar (essentially, individual savings accounts), which can be designed in various ways; · a voluntary third pillar, which is flexible and discretionary (this pillar, too, can take a variety of forms); and · a fourth pillar of informal intrafamily or intergenerational sources of financial and nonfinancial support to the elderly, including access to health care and housing. Pillar Design The design of the Hungarian pension system incorporates all five of these pillars (table 5.2). The publicly managed noncontributory zero pillar, financed with general tax revenues, redistributes income to eld- erly people whose income would otherwise fall below an established minimum level. Both the traditional, publicly managed, pay-as-you-go first pillar and the privately managed, fully funded second pillar are earnings-related schemes. First-pillar benefits are adjusted in retirement using a formula based on the average of inflation and wage growth (Swiss indexation), thereby allowing pensions to rise more rapidly than inflation without imposing the heavier fiscal burden of wage indexation. The second pillar is mandatory. Benefits are a function of an individ- ual's contributions and investment earnings. At retirement, account balances are converted into annuities based on the accumulated capi- tal in an individual's account and the individual's conditional life expectancy. The third pillar is an optional privately managed, fully 152 Table 5.2 Structure of Hungary's Pension System Taxation Investment Generic Benefit income / Scheme type Coverage Type Function Financing benefit Indexation Contributions capital gains Benefits Zero pillar Universal Means tested Redistributive Tax revenues Supplement Based on n.a. n.a. Exempt (public non- to actual old-age contributory) income to minimum reach 80 pension percent of old-age minimum pension First pillar Mandatory Defined Insurance Percentage of Based on 50 percent Taxed n.a. Exempta (public, benefit individual years of inflation, earnings earnings service and 50 percent related) wage history wages Second pillar Mandatory Defined Insurance Percentage of Pension from 50 percent Taxed Exempt Exempta (private, contribution individual conversion inflation, earnings earnings of capital 50 percent related) accumulation wage- into annuities indexed annuity Third pillar Voluntary Defined Insurance Voluntary Pension from Depends Exemptb Exempt Exemptc (private, contribution contributions capital on options voluntary) accumulation chosen Fourth pillar Mandatory n.a. Insurance Percentage of Specified n.a. Exempt n.a. n.a. (public individual health health care) earnings service plus tax package revenues Source: Authors' compilation based on data from European Commission 2007a, 2007b; OECD 2001; and unpublished information provided by World Bank staff. n.a. = Not applicable. a. There is great uncertainty regarding the taxation of benefits after 2013. b. Thirty percent of contributions are tax deductible up to an annual cap of 100,000 forint. c. If the pension is taken as a qualified annuity and the accumulation period is 20 (or more) years, than 10­20 years of accumulation are partially taxed. 153 154 Adequacy of Retirement Income after Pension Reforms funded defined-contribution pension scheme intended to provide individu- als with a mechanism for supplementing the benefits provided by the mandatory pillars. The fourth pillar provides health care to the elderly as part of the national health care system. First-pillar contributions are taxed; benefits are tax exempt until 2013, after which they may be taxed, subject to some allowance.5 Although in 1997 such future taxation policies may have been envisaged by lawmak- ers, there is no current regulation to this effect. The fully funded second pillar is treated similarly: contributions are fully taxed, while returns on investment and benefits are untaxed (unless this is also changed in 2013). The voluntary third pillar broadly follows an exempt-exempt-exempt regime, meaning that contributions are exempt from taxation (subject to a ceiling), investment income is fully exempt, and benefits are exempt when individuals draw a qualified annuity based on 20 or more years of accumulation (see box 1.1 in chapter 1). The noncontributory zero pillar (which provides a means-tested benefit for the poor) and the fourth pillar (which provides health care coverage) are completely tax exempt. Noncontributory scheme. Hungary provides a noncontributory old-age allowance as part of its overall social assistance program to ensure a min- imum level of income for the elderly. To be eligible, applicants need to be at least age 62 and able to demonstrate that their total income--from gainful activity, revenues from investments, and other sources, net of taxes and social insurance contributions--falls below 80 percent (95 percent for couples) of the minimum old-age pension. The allowance is means tested and adjusted in value such that the beneficiary's total income reaches the minimum threshold. In 2003, this allowance was paid to 6,679 beneficiaries (roughly 0.4 percent of the population age 62 and older), at a cost equivalent to 0.01 percent of GDP (World Bank 2007b). Earnings-related schemes. Hungary's earnings-related pension scheme consists of two pillars (table 5.3). The first pillar is a traditional defined-benefit pension scheme financed on a pay-as-you-go basis. The second pillar, introduced in 1998, is a fully funded defined-contribution pension scheme. The reforms of the late 1990s introduced parametric changes to the first pillar. These changes included raising the retirement age (from 60 to 62 for men and from 55 to 62 for women),6 changing the formula by which pensions are calculated, and changing how pensions are adjusted over time for inflation. Table 5.3 Parameters of Earnings-Related Schemes in Hungary before and after Reform Vesting/minimum Contribution Pension assess- Scheme type Period eligibility period Contribution ratea ceiling Benefit rate ment base Retirement age First-pillar Prereform 20 years normally; 30 percent Employee 33 percent for Annual wage 60 for men, 55 for (earnings related) 15 years at age (24 percent by contributions the first 10 years, history since women 62 under special, employer, 6 subject to ceiling; 25 percent per 1988. Since strict conditions percent by no ceiling for year thereafter 2008, annual net employee) employer wages are used; contributions revaluation covers all but last year before retirement Postreform 20 years normally; 25.5 percent Employee: set Until 2013: 33 Average lifetime 62 for everyone as 15 years at age (33.5 percent)a annually by the percent for the earnings valorized of 2009 62 under special, (24 percent by government first 10 years, 2 by wage growth strict conditions employer, (at about eight percent for 11­25 1.5 percent times the mini- years, 1 percent [9.5 percent] mum wage) for 26­36 years, by employee) Employer: no 1.5 percent maximum beyond 36 years After 2013: 1.65 percent per year (continued) 155 156 Table 5.3 Parameters of Earnings-Related Schemes in Hungary before and after Reform (continued) Vesting/minimum Contribution Pension assess- Scheme type Period eligibility period Contribution ratea ceiling Benefit rate ment base Retirement age Second-pillar (earnings related) Prereform n.a. n.a. n.a. n.a. n.a. n.a. Postreform No minimum 8 percent (paid by Employee: Pension from Accumulated 62 for everyone as employee) set annually by capital funds of 2009 the government accumulation (at about eight times the mini- mum wage) Employer: no maximum Source: Authors' compilation based on information from the Ministry of Finance; European Commission 2007a, 2007b; Palmer 2007; and U.S. Social Security Administration 2006. n.a. = Not applicable. a. For individuals participating in the first pillar only, employers pay 24.0 percent and employees pay 9.5 percent, for a total of 33.5 percent. For individuals participating in both the first and the second pillars, employers pay 24.0 percent; employees pay 1.5 percent (in 2008) to the first pillar and 8 percent to the second pillar. The first-pillar contribution rates for employers and employees change almost annually. The second-pillar rate paid by employees was raised from 6 percent in 1998 to 7 percent in 2003 and 8 percent in 2008. Hungary 157 The prereform formula suffered from several key design flaws. First, the credit awarded for years of service fell with each additional year of service. Ten years of service provided income replacement of about 33 percent of preretirement wages, whereas 4 times this length of service resulted in a benefit only 2.5 times higher (Palmer 2007). Second, the pension assessment base was computed using only three of the top five years of service (taking into account only wages earned after 1988), which failed to capture an individual's actual wage his- tory. Third, the cap on earnings used to calculate benefits created incentives for the underdeclaration of earnings.7 The reforms changed all this. The annual accrual rate was fixed at 1.65 percent of the pension assessment base of service starting in 2013, thereby granting equal weight to all years of service. The pension assessment base is based on average valorized wages earned since 1988, thereby strengthening the link between contributions and benefits. The cap on earnings was substantially increased and is currently about three times average earnings. In 2001, Swiss indexation was introduced, in which pensions are adjusted in retirement using a formula based on the average of inflation and wage growth, thereby allowing pensions to rise more rapidly than inflation but without imposing the heavier fiscal burden of wage indexation. The reforms of the late 1990s also established a second-pillar pen- sion scheme. The new scheme is mandatory for new entrants and vol- untary for those with an established history of contributing to the old pay-as-you-go system. Of the total contribution rate of 33.5 percent in 2008, 25.5 percentage points (24.0 percentage points paid by employ- ers and 1.5 percentage points paid by employees) flow to the first pil- lar, while 8.0 percentage points (paid entirely by employees) flow to the second pillar. For people enrolled in both pillars, benefits accrue under both pillars so that the accrual rate under the first pillar is reduced from 1.65 percent a year of service to 1.22 percent (this also applies to years before 1998). The reduced first-pillar benefit will be supplemented by the second pillar as a function of both an individual's account balance at retirement and his or her age at retirement. Voluntary scheme. Hungary introduced a voluntary private pension scheme in 1994 (table 5.4). Benefits are provided after a minimum of 10 years of contributions. Participation is encouraged using tax incentives. Employers are permitted to make contributions on behalf of their employees. The minimum contribution level is established by each individual fund. 158 Adequacy of Retirement Income after Pension Reforms Table 5.4 Characteristics of the Voluntary Scheme in Hungary Lump-sum Tax advan- Contributions payments Vesting tages to tax deductible possible in Coverage period Retirement age participants by employers retirement Every formally 10 years Benefit can be Yes Yes Yes and/or legally withdrawn employed per- after 10 years son, including (no set retire- self-employed ment age) persons Sources: Leonik 2006; OECD 2001; and consultations with World Bank staff. Employer contributions to the voluntary scheme are exempted from the social insurance levy up to 50 percent of the minimum wage. Thirty per- cent of total contributions (including employer contributions) up to an annual cap of 100,000 forint (Ft) are tax deductible, with the amount of the allowance credited to the participant's account (the cap is Ft 130,000 if plan members retire before 2020). Benefits can be collected in the form of a lump-sum payment or as an annuity or be spread over a prespecified period in retirement. In 2006, benefits totaled Ft 29.2 billion. Annuities accounted for a negligible share. In 2007, 69 private pension funds operated in the market, supervised by the Hungarian Financial Supervisory Authority. In 2005, private pen- sion funds had assets totaling Ft 590.11 billion (2.9 percent of GDP) and were receiving contributions from 1,307,222 participants (31 percent of the labor force and 18.8 percent of the working-age population). In 2006, the bulk of investments were in government bonds (Hungarian Financial Supervisory Authority 2007). Since 1994, the number of participants in the voluntary scheme has increased rapidly, but the value of assets accu- mulating in the scheme remains relatively small. Health care system. Health care in Hungary is provided primarily through a mandatory health insurance scheme (the Health Insurance Fund) operated by the National Health Insurance Fund Administration (NHIFA). Health care services are delivered predominantly by public providers owned by local governments through contracts with NHIFA financed by contributions from the covered population. Salaried employ- ees are required to pay a contribution rate of 15 percent (4 percentage points paid by employers and 11 percentage points paid by employees). Self-employed people contribute 15 percent of their declared earnings. Hungary 159 Some people, including pensioners and people with low incomes, are not required to pay contributions. Health care services for such people are financed by the Health Insurance Fund and the government. Recipients of health care are required to make copayments in order to be eligible for some services, pharmaceuticals, and devices. In 2005, health expenditure accounted for 7.8 percent of GDP, 70.8 percent of which was public expenditure and 29.2 percent of which was private expenditure. Of the private expenditure, 86.8 percent was attributable to out-of-pocket expenditure (informal payments, direct payments, and copayments) (WHO 2008). Institutional Structure and Coverage of Earnings-Related Schemes The mandatory pillars of the Hungarian pension system cover salaried employees and self-employed persons. Independent farmers may partici- pate on a voluntary basis. The State Tax Collection Agency collects con- tributions for the first-pillar pension scheme. The Central Administration of National Pension Insurance administers the scheme. In 2005, 3,881,000 people contributed to the scheme, accounting for 92 percent of the labor force and 56 percent of the working-age population (Hungarian Financial Supervisory Authority 2007). The Hungarian Financial Supervisory Authority is responsible for regulating private pension fund companies. In 2007, 20 pension funds were operating in the (mandatory) second-pillar pension fund market. Contributions to the second pillar are deducted from salaries by employers and transferred directly to the pension funds. In 2005, 2,509,941 people contributed to the scheme (roughly 65 percent of the number of partic- ipants in the first-pillar scheme). In 2005, the assets of the second pillar totaled Ft 1,221 billion (5.6 percent of GDP). Structure of Benefits The earnings-related pension scheme provides old-age, disability, and sur- vivorship pensions (from the first pillar only). The provisions governing each of these types of benefits are discussed as follows. Old-age benefits. Beginning in 2013, the accrual rate under the first- pillar pension scheme will be 1.65 percent per year of service for people not enrolled in the second pillar (mostly older workers) and 1.22 percent for people enrolled in the second pillar (who will also receive benefits from their individual funded accounts). To be eligible for a reduced old- age pension under the first-pillar pension scheme, participants must have 160 Adequacy of Retirement Income after Pension Reforms 15 years of service. A full pension requires 20 years of service. Both require participants to have reached the statutory retirement age of 62 for men and 61 for women (in 2009, the ages will be equalized at 62). Until 2009, retirement as early as age 57 is allowed without a reduction in ben- efits provided participants have the required years of service at the time of their retirement. Beginning in 2009, 40 years of service will be required for early retirement, and penalties will be imposed for anyone retiring before age 59. If participants choose to retire early, benefits will be reduced by 1.2 percent per year of early retirement. An accrual rate of 0.5 percent is awarded for each month of service worked after the statu- tory retirement age. Under the second pillar, workers who contributed for less than 15 years may opt to receive the accumulated capital as a lump sum; all other workers must accept an annuity. Unisex mortality tables are used to cal- culate the annuity. Annuities may be provided by the pension fund itself or be purchased from an insurance company. Like first-pillar benefits, annu- ities are indexed 50 percent to inflation and 50 percent to wage growth.8 Disability benefits. Under the first pillar, there are three categories of disabil- ity benefits, depending on the degree of incapacitation: 100 percent loss of capacity to work and the need for permanent care (Category I), 100 percent loss of capacity to work and no need for permanent care (Category II), and at least 67 percent loss of capacity to work (Category III) (table 5.5) Eligibility depends on an individual's age and service at the time of disabil- ity. Benefits depend on the individual's years of service at the time of dis- ability and the degree of incapacitation; they cannot exceed the individual's average earnings. For people in Category III, benefits are equal to 100 per- cent of the old-age pension if the individual has at least 25 years of service Table 5.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in Hungary under the First-Pillar Earnings-Related Scheme Vesting period Contributions Eligibility Benefit rate Partial pension Under age 22: 2 years Included in Loss of at 37.5­100 37.5­63.0 percent Age 22­24: 4 years overall pension least 67 percent of of average indi- Age 25­29: 6 years contribution percent of average individ- vidual earnings Age 30­34: 8 years rate (estimated capacity ual earnings, Age 35­44: 10 years at about to work depending Age 45­54: 15 years 4 percent) on level of dis- Age 55 and over: ability and 20 years years of service Sources: European Commission 2007a; GVG 2003. Hungary 161 and 37.5­63.0 percent of the individual's average earnings if he or she does not. For Category I, the benefits are 100 percent of the Category III bene- fits; benefits for Category II are 105 percent of the Category III benefits (European Commission 2007a). The new disability benefit system that came into effect in January 2008 aims to encourage a return to the labor market by new claimants who have been assessed to have remaining work capacities. These individuals have to participate in a rehabilitation plan designed by the employment office. Any suitable job offer received by the individual from the employment authority has to be accepted. For those eligible for the rehabilitation scheme, a transitory rehabilitation benefit is commensurate to the length of the rehabilitation process, albeit capped at three years. The new rehabilitation system involves a focus on training as a way to strengthen the remaining abilities and skills of disabled individuals. The second pillar provides no disability benefits. In the event of disability, individuals may choose between receiving the accumulated capital in the individual account as a lump sum or having the account balance transferred to the first pillar to improve the disability benefit under that pillar. The benefit under the publicly managed scheme does not depend on the amount transferred (OECD 2001). Survivor benefits. Survivor benefits are awarded to dependents if the deceased had been receiving (or had met the criteria to receive) an old-age or disability pension (table 5.6). Eligible survivors include a current or former spouse, orphans, parents, domestic partners, sisters, brothers, and grandchildren. Provided that the surviving spouse is not already receiving a pension, the survivor benefit is 60 percent of the deceased's pension. If the spouse is receiving a pension, the survivor benefit is 30 percent of the deceased's pension. The survivor benefit is permanent if the spouse is dis- abled, raising at least two minor children, or has already reached retire- ment age. If the spouse was already older than the retirement age when he or she got married, the spouse is entitled to a survivor benefit only if the couple cohabited for at least five years or had a child together. Unmarried couples must have lived together for at least one year if they have a child; otherwise, they must have lived together for at least 10 years. Divorced survivors--or survivors who have been separated for more than one year-- are entitled to a benefit only if they were entitled to alimony. Orphans are entitled to 30 percent of the deceased's pension if one parent dies and 60 percent of the higher of the two parents' pensions if both parents die. Orphans may draw a benefit through age 16, unless enrolled full time as a student, in which case eligibility continues through age 25. Disabled 162 Adequacy of Retirement Income after Pension Reforms Table 5.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in Hungary under the First-Pillar Earnings-Related Scheme Spouse Orphan Total replacement Benefit Remarriage Orphan replacement family Eligibility rate duration test age limit rate benefit Eligibility of 60 percent of For life, if Remarriage 25 30 percent of Cannot deceased deceased's spouse is test before deceased's exceed for old-age pension disabled, is retirement pension; benefit to or disability (30 percent caring for age 60 percent which pension if receiving at least of the deceased own pen- two chil- higher of was enti- sion) dren, or is the two tled above parents' retirement pensions age; 12 or if child 18 months loses both if spouse parents is caring for a child Sources: European Commission 2007a; GVG 2003. orphans may draw a benefit indefinitely. For all categories of survivors, total benefits may not exceed the benefit to which the deceased was orig- inally entitled. Under the second pillar, participants may designate a beneficiary and elect to receive survivor benefits in a lump sum or in the form of annu- ity. Benefits are determined by the balance in the deceased's private pen- sion account at the time of his or her death. Assessment of the Performance of Hungary's Pension System The World Bank has established four principles for evaluating public pension systems, which together should guide the process of pension reform (see Holzmann and Hinz 2005). Broadly speaking, these princi- ples include the adequacy and security of benefits, the affordability of contributions, the sustainability of the system over time, and the robust- ness of the system in the face of demographic changes and macroeco- nomic shocks. This chapter focuses primarily on the adequacy of benefits and the financial sustainability of the earnings-related pension scheme. The remaining principles are mentioned only briefly. Adequacy is ana- lyzed through the lens of net replacement rates. Financial sustainability is evaluated using projections of pension expenditure and revenues. Hungary 163 Benefit Adequacy Replacement rates are a useful yardstick for measuring the adequacy of pension benefits, because they express benefits relative to preretirement earnings, thereby indicating the degree to which income is replaced when workers retire. Two variants are commonly used. Gross replacement rates compute income replacement as the ratio of benefits paid to pretax prere- tirement earnings. Net replacement rates compute income replacement as the ratio of benefits received (that is, after the payment of taxes and other levies, including contributions for social insurance) to posttax preretirement earnings. In general, net replacement rates are a more useful measure of benefit adequacy, because they capture the degree to which actual take- home pay is replaced when workers retire. The level of income replacement at retirement is not the only measure of benefit adequacy. For a full assessment of benefit adequacy, it is also important to determine how postretirement indexation rules will affect replacement rates during retirement. Pension benefits in retirement are expected to be indexed to inflation, so that their real value is maintained. In a growing economy with rising real wages, however, mere price indexa- tion of pensions leads to a deterioration of the relative consumption posi- tion of the retirees. For this reason, some countries have introduced mixed indexation of pensions that use varying weights of inflation and wage growth in the indexation formula. For an evaluation of the effect of indexation on replacement rates in Hungary, the replacement rates are normalized to 100 percent and the assumptions for calculating the replacement rates are maintained (that is, inflation is 2.5 percent a year and real wage growth is 2.0 percent a year). The change in the replacement rate is measured in comparison with full wage indexation or the earnings of an active worker. The results of this analysis indicate that the relative income position of a retiree would dete- riorate by 7 percent after 10 years in retirement and by 21 percent after 35 years in retirement as a result of 50 percent inflation and 50 percent wage indexation of benefits in both the first and the second pillars. The evaluation of income replacement that follows considers replacement rates only at retirement; it does not take into account the impact of index- ation policies on replacement rates during retirement. Replacement rates are a function of the formula governing pension benefits; an individual's contribution history; and, in the case of net replacement rates, the rules of income tax, social security contribu- tions, and other relevant levies. The benefit formula establishes the degree to which the system redistributes income across individuals of 164 Adequacy of Retirement Income after Pension Reforms different levels of preretirement earnings. Progressive systems provide higher levels of income replacement to people with lower levels of preretirement income. In general, the degree to which a system is redistributive depends on the existence (and value) of flat transfers and minimum pension guarantees, the degree to which benefits are earn- ings related, and the existence of ceilings on earnings subject to contri- butions. An individual's contribution history can be characterized by his or her age of entry into the labor force, contribution density, and decisions regarding the timing of retirement. To some degree, these three factors are influenced by the incentives embodied in the pension system. The tax and contribution system influences net replacement rates through the progressiveness of the income tax formula, which taxes (higher) income during a worker's active life more so than it does (lower) pension benefits in retirement. In addition, there are social security levies (for pensions; unemployment; health care; and, at times, housing and family benefits), which are typically reduced or eliminated altogether in retirement. These benefits are particularly important for low- to middle-income groups. Benchmarks need to be established for the evaluation of the adequacy of the income replacement provided by the earnings-related pension schemes. Unfortunately, there is no consensus on what constitutes ade- quacy. According to one widely respected definition, pensions are ade- quate when they are sufficient to prevent poverty among the elderly and provide the vast majority of the population with a reliable mechanism for smoothing income over their lifetime. Even with a definition, however, establishing benchmarks is problematic, because attitudes vary across countries as a result of social and cultural perceptions. Moreover, bench- marks ignore the other factors that affect the welfare of the elderly--and that also vary across countries--including the existence and generosity of health insurance and long-term care, the cost of housing, the structure of traditional living arrangements, the presence of informal intrafamily or intergenerational sources of financial and nonfinancial support, and the availability and security of other mechanisms for people to save for their own retirement. One reputable nine-country study (OECD 2001) observes that living standards are roughly comparable for people 10 years older than the normal retirement age and people 15 years younger than the normal retirement age when retirees have disposable income equal to roughly 80 percent of the disposable income of working-age people. In part, this is attributable to the fact that retirees have no work-related expenses (they do not have to commute or buy special clothing or uniforms, for Hungary 165 example). This finding, however, does not imply that mandatory first- pillar pension schemes should actually target an 80 percent net replace- ment rate. To the contrary, in middle- and high-income countries, one can reasonably expect individuals to save for their own retirement--and the empirical evidence suggests that, in practice, they do so.9 There is also some evidence to suggest that the ratio between pre- and postretirement income is somewhat independent of the income replacement mandate of the public pension system. Put simply, individuals tend to save more in countries with more modest mandates (and vice versa). Because Hungary has access to relatively well-developed financial markets, it would seem reasonable to expect middle- and high-income workers to save enough to finance at least 25 percent, if not closer to 50 percent, of this 80 percent income replacement target. Given this, three benchmarks are provided: a 40 percent net replacement rate (which implies that individuals would be expected to save enough to finance half of the total income-replacement target); a 60 percent net replace- ment rate (which implies that individuals would be expected to finance a quarter of the target); and an 80 percent net replacement rate (which implies that individuals, most of whom would be low-income earners, would not be expected to contribute anything toward the target).10 In the following analysis, these three benchmarks are used to evaluate the adequacy of benefits in Hungary in conjunction with the average net replacement rate observed in 53 countries, the average net replacement rate observed for selected countries in Europe and Central Asia, and the poverty line in Hungary. To estimate gross and net replacement rates, we consider two criti- cal dimensions--earnings levels and contribution periods--with the help of the Analysis of Pension Entitlements across Countries (APEX) model.11 This model generates estimates for replacement rates under steady-state assumptions (that is, as if the rules of the reformed pen- sion scheme had been in place over the entire active life of the individ- ual). Because life expectancies at retirement are projected to increase over time--which will affect the benefits paid by defined-contribution pension schemes--a reference year must be chosen. For the purpose of this study, 2040 is used because it provides a sufficiently long contri- bution period over which to approximate steady-state conditions. The first critical task is to investigate levels of income replacement across a relevant spectrum of income. Income is represented as a percent- age (50­200 percent) of average earnings. The second task is to investigate the impact on income replacement of differences in the duration, timing, and density of an individual's contribution history (density refers to the 166 Adequacy of Retirement Income after Pension Reforms percentage of time an individual actually contributes over a given period). To facilitate the presentation of these multidimensional results, we com- pute replacement rates as a function of the age at which an individual exits the labor market. They are presented separately for full-career and partial- career workers. Replacement rates for full-career workers. Projected replacement rates for full-career workers in 2040 are examined first. For this analysis, a full career is defined as continuous employment from age 20 to the current normal retirement age of 62. Gross replacements rates increase as income levels fall (figure 5.3). The share of second-pillar benefits is constant (at 26.2 percentage points) for all income levels, irrespective of income. The amount of first-pillar bene- fit increases as income levels decline, illuminating the relatively redistrib- utive feature of the first pillar.12 The situation changes when taxes are taken into consideration. As a result of the impact of taxes and contributions, middle- and high-income workers receive higher net replacement rates than do low-income workers (figure 5.4).13 This result reflects the nature of the tax code in Hungary, not the design of the pension system. Pensions for full-career workers in Hungary are adequate (figure 5.5).14 Replacement rates for all levels of preretirement income are higher than their respective benchmarks, suggesting that the pension system is Figure 5.3 Sources of Gross Replacement Rates in Hungary, by Income Level 80 70 60 (percent) 50 rate 40 30 replacement 20 10 gross 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (defined benefit) Sources: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. Hungary 167 Figure 5.4 Sources of Net Replacement Rates in Hungary, by Income Level 100 90 80 70 (percent) 60 rate 50 40 30 20 replacement 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (defined benefit) taxes Sources: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. Figure 5.5 Net Replacement Rates for Male Full-Career Workers in Hungary, Europe and Central Asia, and the World, by Income Level 100 80 high benchmark (percent) rate 60 middle benchmark 40 low benchmark replacement net 20 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage Hungary average Europe and Central Asia average world average poverty line (percentage of average income) Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. 168 Adequacy of Retirement Income after Pension Reforms effectively smoothing consumption from work into retirement.15 Indeed, even middle- and high-income individuals are receiving levels of income replacement above the highest of the three benchmarks. The downside of such effective income smoothing is that the system provides little incen- tive for workers to save outside of the pension system. Given that bene- fits for even the lowest full-career workers greatly exceed the poverty line, the objective of poverty alleviation is being met. Levels of income replacement in Hungary are higher than the regional and world averages, especially for high-income workers. Replacement rates for partial-career workers. Not everyone works from age 20 to the statutory retirement age. Many individuals enter and exit the labor force (often at different ages and for different periods of time) and earn different wages while working. To examine the adequacy of ben- efits for partial-career workers, we consider three stylized cases. These cases include career type A (someone entering the labor force at age 25 who works continuously for a period of years before leaving the work- force at some point between the ages of 50 and 70 and then claims a ben- efit); career type B (identical to career type A, except that the worker enters the workforce at age 30 and leaves no earlier than age 55); and career type C (identical to career type A, except that the individual con- tributes in only three years out of four while in the labor force). In cases in which the withdrawal from the formal labor market occurs before the statutory retirement age, the pension is claimed (and the replacement rate calculated) only at the later age. For withdrawals after the statutory retire- ment age, the ages coincide. Because a minimum of 20 years of contributions are required for benefit eligibility, the minimum exit ages for the three career types vary (figure 5.6). As a result, the lines for the replacement rates start at different points. The steep rise in replacement rates at age 62 reflects the fact that there is no longer an actuarial reduction in ben- efits for early retirement; each additional year of contributions con- tributes directly to higher benefits. Several conclusions can be drawn from figure 5.6. First, entering the workforce later in life is costly (someone entering the workforce at age 30 receives a net replacement rate 8­19 percentage points lower than someone entering the workforce at age 25). Second, working intermit- tently is costly (someone who enters the workforce at a given age but who contributes only three years out of four will receive a net replace- ment rate 15­40 percentage points lower than someone who contributes continuously). Third, the Hungarian pension system provides partial-career Hungary 169 Figure 5.6 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Hungary, by Career Type and Exit Age 150 140 130 120 110 100 (percent) 90 high benchmark rate 80 world and Europe and Central Asia average 70 middle benchmark 60 50 low benchmark 40 poverty line (percentage of average income) replacement 30 20 net 10 0 45 50 55 60 65 70 exit age from labor market career type A: career type B: career type C: entry age--25, entry age--30, entry age--25, contribution contribution contribution density--100% density--100% density--75% Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. See text for descriptions of career types. workers with an acceptable level of income replacement, as evidenced by the fact that even people who exit the labor market long before reaching the statutory retirement age receive a net replacement rate higher than the poverty line. People who work until the statutory retire- ment age (which is likely to include people with limited ability to save on their own, for whom the mandatory system is likely to be their main source of retirement income) receive a net replacement rate of 60 per- cent or higher (the middle of the three benchmarks). To receive a replacement rate of 80 percent (the highest of the three benchmarks), career type A workers need to work until retirement age, career type B workers need to work two years beyond retirement age, and career type C workers need to work well beyond the normal retirement age. Fiscal Sustainability The sustainability of a pay-as-you-go, first-pillar pension scheme is best evaluated in actuarial terms by estimating the scheme's actuarial deficit as the difference between its assets and liabilities. If a large actuarial deficit exists, the scheme is financially unsustainable and needs policy actions that increase its assets, reduce its liabilities, or both. A good proxy 170 Adequacy of Retirement Income after Pension Reforms for the actuarial deficit is the difference between the present value of the scheme's expected future revenues (that is, contributions and other income) and the expected future expenditures (that is, benefit payments, administrative costs, and other expenses) over an extended projection period. The difference between these two values represents an unfunded liability (sometimes referred to as a financing gap) on the public-sector balance sheet. Because this study is also concerned with the time path of revenues and expenditures (and the resulting balance across the projec- tion period ending in 2050), this more pragmatic approach has been taken. Projections of expenditures, revenues, and deficits are presented on the basis of available postreform fiscal projections. The reforms of the late 1990s considerably improved the sustainability of the Hungarian pension system. The revenues of the first pillar are projected to hover at roughly 6.5 percent of GDP between 2005 and 2050, while expenditure is estimated to rise to 9.4 percent of GDP, resulting in a deficit of 2.9 percent of GDP--roughly half what had been projected.16 Despite this significant improvement, the first-pillar scheme will continue to gener- ate substantial deficits even after the reforms have been fully implemented (figure 5.7). This is partly because of the transition deficit and the diversion of contribution revenue from the first pillar toward the second pillar.17 Figure 5.7 Projected Fiscal Balance of Hungary's Public Pension System after Reform, 2005­50 10 8 6 GDP of 4 2 percentage 0 ­2 ­4 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 year expenditures revenues deficit Source: Orbán and Palotai 2005. Note: Projections were revised in 2007. Hungary 171 What options exist for restoring the system to fiscal balance? Unfortunately, for policy makers, the options are limited. At the level of economic policy, an increase in employment rates across all ages would clearly help, but it would not resolve issues that make the scheme unsustainable. Although higher contribution revenues (resulting from increased employment) will reduce cash deficits, they will also create liabilities that will eventually become due. If an unfunded system is actu- arially unsustainable because benefit promises per contributor exceed contribution payments revalued by a sustainable (implicit) rate of return, then greater employment offers a temporary cash respite while actually making the system more unsustainable. This leaves policy makers with three basic options: increasing contri- bution rates, reducing benefits, or raising the retirement age. Revenues can be increased by raising the contribution rate (or by transferring funds from general revenue to compensate partly or fully for the tran- sition deficit). Alternatively (or in addition to, because the options are not mutually exclusive), expenditures can be reduced by cutting ben- efits, increasing the number of years required to become eligible for benefits, or delaying the payment of benefits by raising the retirement age further. Because raising the contribution rate could threaten com- petitiveness and will likely strengthen incentives for tax evasion, it is typ- ically not embraced (it would also represent a reversal of policy because, until recently, Hungary had been reducing the contribution rate to reduce the adverse impact of high taxes on labor markets). Steps such as raising retirement ages and tightening eligibility criteria have already been taken; additional increases in retirement ages and further tightening of eligibility criteria will be needed to make the pension sys- tem sustainable. An informal analysis suggests that retirement ages for men and women would have to be increased to at least 73 by 2050 to restore long-term fiscal balance.18 If retirement ages are left unchanged and the current structure of the system is retained, further cuts in benefits--on the order of 41 per- cent from the public pillar--would be required for the system to become sustainable. Cuts of such a magnitude raise the question of whether restoring sustainability would not exact a cost in terms of the adequacy of benefits provided to future beneficiaries. If benefits are adjusted to maintain a similar fiscal balance in proportion to the overall size of the first-pillar scheme, full-career workers will receive replacement rates roughly 20 percentage points lower in 2050 than they receive today (figure 5.8). 172 Adequacy of Retirement Income after Pension Reforms Figure 5.8 Net Replacement Rates for Male Workers in Hungary before and after Benefit Adjustment 100 80 high benchmark (percent) rate 60 middle benchmark 40 low benchmark replacement net 20 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage net replacement rate adjusted net replacement rate poverty line (percentage of average income) Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. Two conclusions emerge from comparing these new (lower) net replacement rates against the three benchmarks. First, a 41 percent reduc- tion in benefits does not cause the income replacement for middle- and high-income full-career workers to fall anywhere near the poverty line. Replacement rates for the lowest-income workers are only slightly below the poverty line. This indicates that a sustainable first-pillar pension scheme in Hungary would still achieve its poverty alleviation objective for almost all workers. Second, the reduction in benefits will still support the objective of smoothing lifetime consumption for middle- and high-income full-career workers, because levels of income replacement are higher than the 80 per- cent benchmark. This objective may not be met for low-income work- ers, however, for whom levels of income replacement are lower than the 80 percent benchmark. This observation is subject to two caveats. One is that this analysis considers only full-career workers, while the average worker now contributes for only about 27­30 years, substantially less than the 40 years expected over a full career. Contributing to the pension Hungary 173 scheme for only 30 years, for example, reduces net income replacement by 27 percentage points for middle-income workers. Another caveat is that workers always have the option of saving outside the first-pillar pen- sion scheme. To increase income replacement by 1.0 percentage point, for example, a full-career worker would need to save only about 0.5 percent of his or her earnings from age 40 to the current age of retirement.19 Conclusions In response to emerging deficits in the existing, pay-as-you-go, public pen- sion system (and projections that deficits would grow substantially over the long term as a result of the aging of the population), Hungary launched an ambitious program of reform in the mid-1990s. That program intro- duced parametric changes to the existing defined-benefit scheme and shifted (in 1998) the system from a monopillar design toward a multipil- lar design that included a mandatory fully funded defined-contribution scheme (a voluntary third-pillar scheme had been introduced in 1994). These reforms were considered radical, because Hungary was the first country in the region to attempt to finance a portion of pension benefits on the basis of funded defined-contribution accounts. As a result of these reforms, the long-term fiscal position of Hungary's pension system is projected to improve substantially, with projected deficits in 2050 falling from 6.0 percent of GDP to 2.9 percent of GDP despite the additional burden of the transition deficit through the intro- duction of the funded second pillar. Some postreform adjustments to the first-pillar scheme, such as the introduction of the 13th monthly pay- ment, have made the system less sustainable; more recent reforms, in par- ticular the move toward a net-income pension base, have moved the system back toward sustainability. Overall, the reforms have broadly strength- ened the link between preretirement contributions and postretirement benefits. They are, therefore, expected to create incentives for people to remain in the workforce and contribute for longer periods. The resulting gross and net replacement rates for full-career workers in Hungary are projected to be well above regional and international aver- ages (see chapter 1). The future net replacement rates for full-career workers are projected to exceed 80 percent (and to approach 100 percent for some workers) for the analyzed income spectrum. As in other coun- tries, partial-career workers--workers who left the workforce before reaching the retirement age, worked intermittently, or have gaps in their employment history--risk receiving a level of income replacement closer 174 Adequacy of Retirement Income after Pension Reforms to or lower than the 40 percent benchmark. Low contribution density is of substantial concern for benefit adequacy if recent behavior is repeated in the future. Over the period 1997­2006, 43 percent of those employed worked less than eight months a year and 28 percent worked less than six months a year. Improving formal labor force participation and employ- ment rates across all age groups is clearly a key economic policy task for the years come. High replacement rates that reflect more recent changes may be a reaction to concerns about the "lost generation" emerging from transition. Many members of this generation will approach retirement in the next 10­20 years. These workers have low earnings histories and will receive low benefits if they qualify for pensions at all. Although handling their income needs through changes in the benefit formula may be tempting and politically expedient, a long-term perspective on benefit adequacy in financial sustainability of the reform schemes would suggest tailored and transitory measures. Permanent changes to address the needs of the lost generation will undermine not just adequacy and sustainability but also the contribution­benefit link and will encourage continued informality for the future. Replacement rates have been computed under the assumption that funded pension schemes earn a rate of return of 1.5 percentage points more than wage growth. This earnings differential broadly reflects the per- formance of pension funds in Organisation for Economic Co-operation and Development countries over the past 30 years. The earnings differen- tial in emerging economies is almost twice as large (Holzmann 2009). The performance of pension funds in Hungary since their inception, however, has been well below this benchmark. If such performance continues, the Hungarian pension system will not be capable of delivering the replace- ment rates projected earlier.This concern calls for a review of pension fund performance and accelerated progress in financial market development. To restore long-term fiscal balance to the pension scheme without changing the benefit design, policy makers need to raise retirement ages, possibly to 73 by 2050. To achieve the full fiscal impact of this measure for the public pillar, they would need to fix the replacement rate at the current retirement age. Increasing the retirement age in step with increases in life expectancy at retirement is a natural choice, both for indi- viduals and for policy makers, but it requires cross-sectoral policy reforms to enable elderly workers to continue to participate in the labor market.20 Another option for restoring fiscal balance is to cut benefits at retire- ment or reduce the generosity of benefit indexation. Rough estimates Hungary 175 suggest that, on average, replacement rates at retirement would have to fall by about 20 percentage points to achieve fiscal balance by 2050. Moving to full price indexation would reduce average benefits by some 10 percentage points. Reducing benefits by this amount would not com- promise the objective of smoothing lifetime consumption for full-career workers of all income levels when measured against the highest of the three benchmarks employed in this analysis or the objective of alleviat- ing poverty among the elderly because levels of income replacement would still not fall anywhere near the poverty line. Individuals who wish to defer more of their lifetime consumption into retirement would still have the option of participating in Hungary's vol- untary third-pillar pension scheme. Given the low per capita payments into the third-pillar pension scheme, however, this would require some major changes in individual savings behavior (and the capacity to do so) to become effective. Voluntary contributions are currently low, however. Major changes in individual savings behavior (and the capacity to save) will be necessary if the third pillar is to become a significant source of retirement income. Notes 1. Unemployment rose from 0.3 percent in 1990 to 10 percent in 1995, while employment rates fell from 76 percent to 58 percent. During this period, roughly 30 percent of formal sector jobs were lost (Palmer 2007). 2. The Hungarian population is projected to age less than most other Organi- sation for Economic Co-operation and Development (OECD) countries, because projected life expectancies for men and women by 2040 are lower than in all OECD countries except Turkey (Kovac 2008). 3. For more information regarding these population projections, see Reiterer (2008). 4. More recent projections of the system dependency ratio, which reflect the impact of the reforms since 1998, suggest a much more optimistic outcome. Projections from the 2008 Pension Round Table indicate that the system dependency ratio will fall below the old-age dependency ratio by 2012 and will remain there until the end of the projection period (2050), at which point the difference between them will reach almost 8 percentage points. Even under these projections, however, the system dependency ratio will increase, from just below 30 percent in 2012 to well over 50 percent by 2050 (conversation with Erzsebet Kovacs). 5. A great deal of uncertainty surrounds taxation policies after 2013. 176 Adequacy of Retirement Income after Pension Reforms 6. Low retirement ages relative to life expectancy contribute to fiscal imbal- ance, because they increase the value of lifetime benefits relative to con- tributions (for individual contributors) and increase the total number of beneficiaries relative to contributors (for the overall pension scheme). 7. As a result of high inflation in 1992­96, the ratio of the cap on earnings used in the calculation of benefits to average gross wages declined from 3.4 to 1.6 (Simonovits 2002). 8. There is an ongoing discussion in Hungary about whether the markets will be able to deliver Swiss indexation as written into the law. Regulations gov- erning annuity provisions are still pending and are expected in late 2008. For a discussion of these issues, see Párniczky (2005). 9. In Chile, for instance, 70 percent of retirees from the mandatory public pension system own their home, which is a form of savings (see Valdés- Prieto 2008). 10. These benchmarks approximate the standards developed by the Inter- national Labour Organization (ILO) (1952) and the Council of Europe (1990). ILO Convention 102 of 1952 sets a minimum benefit equal to 40 percent of the reference wage for married men of pensionable age. This amount was raised to 45 percent in 1968. The European Code of Security of 1990 sets a minimum standard for members of the Council of Europe equal to 65 percent for married people of a specific age. 11. The APEX model was developed by Axia Economics, with funding from the Organisation for Economic Co-operation and Development and the World Bank. The model codes detailed eligibility and benefit rules for first- and second-pillar schemes based on available public information that has been verified by country contacts. Because the details of the rules sometimes change on short notice (and limited public disclosure), the calculations pre- sented here should be considered as best approximations only. 12. The pension is calculated on the basis of the net pension assessment base. Because of the progressive tax system, gross replacement rates increase with increases in income. Because the first pillar involves 13 "monthly" pension benefit payments a year, the additional benefit is apportioned to calculate the replacement rates provided under the defined-benefit scheme. This extra monthly benefit was introduced only recently. This benefit represents a departure from the social insurance principle. Effective January 2009, early retirees will not be eligible for this benefit, and a cap will be intro- duced limiting it to the average wage. 13. Estimates are based on the assumption that real wage growth is 2 percent, the net real rate of return on invested assets is 3.5 percent, and benefits (from the unfunded and funded pillars) are price indexed. As a proxy for the poverty line, a figure of 35 percent of the average net wage is used, because this percentage broadly approximates a US$2.25-a-day poverty Hungary 177 line converted into national currency, adjusted for purchasing power parity, expressed relative to the national average net wage, and averaged across the nine study countries. Such an approach enables valid comparisons to be made across the sample (see chapter 1). 14. Replacement rates are simulated for an unmarried male working a hypo- thetical career path under the assumption that real wage growth is 2 per- cent, inflation is 2.5 percent, the rate of return on invested assets is 3.5 percent, and the worker retires at the statutory retirement age. Replacement rates shown do not consider the benefits received from occupational schemes. 15. There is a slight increase in net replacement rates as income levels increase, as a result of the taxation of earnings and pensions. Otherwise, gross replace- ment rates are equal across income levels. 16. Projections by Orbán and Palotai (2005), covering old-age pensions only, were used for this analysis, because the European Union publishes only those expenditures related to the public pension system. More detailed pro- jections are being developed and are expected to become available in late 2008. These projections are being coordinated by the Pension and Old-Age Round Table, an independent expert body commissioned by Hungary's prime minister. Preliminary results suggest that the long-term deficit may be higher than previously projected. 17. This projected deficit is well below that of the harmonized projection by the Economic Policy Committee­Ageing Working Group (2007). Although pro- jected revenues are roughly similar across the projection period, expenditures for the public scheme are projected to increase to 14.7 percent of GDP, resulting in a deficit of 7.8 percent of GDP by 2050. 18. This estimate is based on the World Bank's baseline demographic projec- tions and assumes that everyone over age 66 receives a pension, everyone age 20­66 contributes, and all pensioners receive the replacement rate awarded to the median worker (40 percent). 19. This estimate is based on the assumption that real wage growth is 2.0 per- cent, the net real rate of return on invested assets is 3.5 percent, and benefits (from both the unfunded and the funded pillars) are price indexed. Country- specific mortality rates are used for this analysis. 20. See Holzmann, MacKellar, and Repansek (2009) for a conference volume that addresses these issues for the countries of southeastern Europe. Bibliography Augusztinovics, M., and J. Köllő. 2009. "Decreased Employment and Pensions." In Pension Reform in South-Eastern Europe: Linking to Labor and Financial Market Reforms, ed. R. Holzmann, L. MacKellar, and J. Repansek, 89­104. Washington, DC: World Bank. 178 Adequacy of Retirement Income after Pension Reforms Council of Europe. 1990. European Code of Social Security (Revised). Rome. European Commission. 2006. "The Impact of Ageing on Public Expenditure: Projections for the EU-25 Member States on Pensions, Health Care, Long- Term Care, Education, and Unemployment Transfers (2004­2050)." In European Economy Special Report 1, Economic Policy Committee, Brussels. ------. 2007a. Mutual Information System and Social Protection (MISSOC) database. http://ec.europa.eu/employment_social/. ------. 2007b. Pension Schemes and Projection Models in EU-25 Member Countries. European Economy Occasional Paper 37, Economic Policy Committee and Directorate General for Economic and Financial Affairs, Brussels. Gaal, P. 2004. Health Care Systems in Transition: Hungary. Geneva: World Health Organization. GVG (Gesellschaft für Versicherungswissenschaft und -gestaltung) and European Commission. 2003. Study on the Social Protection Systems in the 13 Applicant Countries: Hungary Country Report. Cologne. Holzmann, R., ed. 2009. Aging Population, Pension Funds, and Financial Market: Regional Perspectives and Global Challenges for Central, Eastern and Southern Europe. Washington, DC: World Bank. Holzmann, R., and R. Hinz. 2005. Old-Age Income Support in the 21st Century. Washington, DC: World Bank. Holzmann, R., L. MacKellar, and J. Repansek, eds. 2009. Pension Reform in South- eastern Europe: Linking to Labor and Financial Market Reforms. Washington, DC: World Bank. Hungarian Financial Supervisory Authority. 2007. Report on the Development of Supervised Sectors in 2006: Achievements and Risks. Budapest. ILO (International Labour Organization). n.d. Database. http://laborsta.ilo.org/. ------. 1952. ILO Convention 102. Geneva: ILO. ------. 1967. ILO Convention 128. Geneva: ILO. Kovac, E. 2008. "International Comparison of Pension Parameters." Biztositasi Szemle 3: 26­37. Leonik, A. 2006. "Pension Security System in Poland." http://www.seminar-burzy.cz/ files/leonik03-06.pdf. OECD (Organisation for Economic Co-operation and Development). 2001. Insurance and Private Pensions Compendium for Emerging Economies: Private Pensions: Selected Country Profiles. Working Party on Private Pensions Secretariat, Paris. ------. 2005. Pensions at a Glance. Paris: OECD. ------. n.d. OECD Statistical Extracts. http://stats.oecd.org/WBOS/. Hungary 179 Orbán, G., and D. Palotai. 2005. The Sustainability of the Hungarian Pension System: A Reassessment. Magyar Nemzeti Bank, Budapest. Palacios, R., and R. Rocha. 1998. "The Hungarian Pension System in Transition." World Bank, Washington, DC. Palmer, E. 2007. "Pension Reform and the Development of Pension Systems: An Evaluation of World Bank Assistance: Background Paper, Hungary Case Study." World Bank, Washington, DC. Párniczky, T. 2005. "Financial Sector Assessment Program Update: Hungary." Technical Note: Pension--Competition and Performance in the Hungarian Second Pillar, December. World Bank Financial Sector Vice-Presidency, Europe & Central Asia Region Vice-Presidency, and International Monetary Fund, Monetary and Financial Systems Department, Washington, DC. Reiterer, A. 2008. Population Development and Age Structure in Southeastern Europe until 2050. World Bank, Washington, DC. Simonovits, A. 2002. Hungarian Pension System: The Permanent Reform. Institute of Economics, Hungarian Academy of Sciences, Budapest. U.S. Social Security Administration. 2006. Social Security Systems throughout the World: Europe. Washington, DC: Social Security Administration. Valdés-Prieto, S. 2008. Designs for the First-Pillar Pensions and the 2008 Chilean Reform. http://editorialexpress.com/cgi-bin/conference/download.cgi?db_ name=SECHI2008&paper_id=130. Whitehouse, E. 1999. Tax Treatment of Funded Pensions. Washington, DC: World Bank. World Bank. 2005. Pensions in the Middle East and North Africa. Washington, DC: World Bank. ------. 2007a. Pensions Panorama. Washington, DC: World Bank. ------. 2007b. Social Assistance in Central Europe and the Baltic States. Washington, DC: World Bank. ------. 2007c. Pension System and Reform Discussion Paper. HDNSP, Washington, DC. WHO (World Health Organization). 2008. Database. http//:www.who.int/ research/en/. C H A P T E R 6 Poland Poland inherited a public pension system financed on a pay-as-you-go basis (meaning that contributions from current workers are used to pay benefits to current beneficiaries). Like the systems of many of its Eastern European neighbors, Poland's pension system was strained by the country's transition to a market economy, a period characterized by massive economic restruc- turing and a marked drop in formal sector employment. Pension revenues fell while expenditures remained roughly constant, creating fiscal imbal- ance in the pension system. By 1992, expenditures had reached the equiv- alent of 16.3 percent of gross domestic product (GDP), while revenues had fallen to 12.1 percent of GDP, resulting in a deficit of 4.2 percent of GDP. The shortfall had to be financed with general tax revenues. Deficits contin- ued through 1999, although their magnitude varied considerably, partly in response to ad hoc measures taken by the government. Long-term projec- tions suggested that the aging of the population would eventually drive the deficits of the pension system to 6.25 percent of GDP. Among policy mak- ers and social security professionals, consensus emerged regarding the need for reform. Following a long debate, in 1999 the Polish government introduced a new multipillar pension system.1 The existing traditional pay-as-you- go public pension scheme was replaced with a mandatory notional 181 182 Adequacy of Retirement Income after Pension Reforms defined-contribution (NDC) scheme,2 a mandatory, privately managed, defined-contribution scheme, and a voluntary, fully funded, defined- contribution scheme intended to provide workers with a mechanism for saving outside of the mandated schemes. Together, these reforms con- siderably improved the fiscal position of the pension system to the point where the system is expected to generate a surplus of about 1 percent of GDP in the long term. Against this backdrop, this chapter evaluates the Polish pension sys- tem, focusing on fiscal sustainability and benefit adequacy. Adequacy is evaluated through the lens of statutory net replacement rates for differ- ent retirement ages, patterns of contributions, and income levels relative to international benchmarks. This chapter is organized as follows. The next section discusses the motivation for the reforms. The following section describes the key char- acteristics of the reformed pension system. The third section assesses the adequacy of pension benefits and the fiscal sustainability of the system. The last section draws conclusions. Motivation for Reform During Poland's transition from a centrally planned economy to a market economy, the pension system began to experience difficulties meeting its benefit obligations on the basis of the contributions it collected from cur- rent workers. Revenues declined as formal sector employment fell; expen- diture rose as a result of low retirement ages, liberal provisions that permitted workers in many sectors to retire early, and relatively generous benefits.3 In 1992, expenditures reached an amount equivalent to 16.1 percent of GDP while revenues fell to 12.1 percent of GDP, resulting in a deficit of 4.2 percent of GDP, which had to be funded from the general budget. The next few years saw some improvement in the fiscal condition of the pension system as a result of ad hoc measures undertaken by the government. By 1999, the fiscal deficit was 2.2 percent of GDP (table 6.1). As a result of these ad hoc measures, the fiscal condition of the pen- sion system was expected to improve to the point where the system would run a slight surplus by the early 2000s. Over the long term, how- ever, as the impact of these measures diminishes and the population ages, the deficits were expected to return, reaching an amount equivalent to 6.25 percent of GDP by 2050 (figure 6.1). These long-term projections are driven primarily by the aging of the Polish population, reflected in old-age dependency ratios (defined as the population age 65 and older divided by the population age 20­64), Poland 183 Table 6.1 Fiscal Balance of Poland's Pension System before Reform, 1992­99 Year Revenues Expenditures Balance 1992 12.1 16.3 ­4.2 1993 11.8 15.8 ­4.0 1994 12.2 16.1 ­3.9 1995 11.6 12.5 ­0.9 1996 11.9 13.0 ­1.1 1997 11.7 13.7 ­2.0 1998 11.3 12.8 ­1.5 1999 10.4 12.6 ­2.2 Source: Authors' compilation based on information provided by the Ministry of Social Policy. Figure 6.1 Projected Fiscal Balance of Poland's Public Pension System before Reform, 2000­50 20 15 10 GDP of 5 percentage 0 ­5 ­10 2000 2010 2020 2030 2040 2050 expenditures revenues balance Source: Authors' compilation based on information provided by the Ministry of Social Policy. Note: Projections cover revenues and expenses associated with old-age pensions only. which are projected to rise from 21.0 percent in 2005 to 55.3 percent by 2050. The system dependency ratio (defined as the number of peo- ple receiving pensions divided by the number of people contributing to the pension scheme) was projected to rise from 29.2 percent in 2000 to 71.6 percent by 2050, in the absence of reform (figure 6.2).4 To address these problems, in 1999 the Polish government introduced a multipillar pension system consisting of a mandatory NDC pension 184 Adequacy of Retirement Income after Pension Reforms Figure 6.2 Projected Old-Age and System Dependency Ratios in Poland, 2005­50 80 70 60 50 40 (percent) 30 ratio 20 10 0 2005 2010 2020 2030 2040 2050 year system dependency ratio old-age dependency ratio Source: Reiterer 2008. scheme; a mandatory, privately managed, defined-contribution scheme; and a voluntary, privately managed, defined-contribution scheme. This multipillar design was intended to improve old-age income security by diversifying retirement savings. Productivity growth in the Polish labor market and rates of return on invested capital now play equally important roles, with productivity growth driving returns from the pay-as-you-go first pillar and rates of return driving returns from the funded second and third pillars. Together, these reforms are intended to restore fiscal sustain- ability to a pension system that would otherwise have continued to require transfers from the state budget (see Chlon, Gora, and Rutkowski 1999). Characteristics of Poland's Pension System This section describes the main characteristics of Poland's pension system. These characteristics include the design of the individual pillars of social insurance; the rules governing pension system taxation, institu- tional structure, and coverage; and the provisions governing old-age, dis- ability, and survivorship pensions. The design of the pension system is assessed using a conceptual framework developed by the World Bank, which generally recommends including a funded component if condi- tions are appropriate but increasingly recognizes that a range of choices is available to policy makers to provide effective old-age protection in a manner that is fiscally responsible (see Holzmann and Hinz 2005). Poland 185 In general, the World Bank supports pension systems composed of some combination of five basic pillars: · a noncontributory (or zero) pillar (in the form of a demogrant, social pension, or social assistance benefit) intended to provide a minimal level of income protection; · a first-pillar contributory system linked to earnings, which seeks to replace a portion of preretirement income; · a mandatory second pillar (essentially, individual savings accounts), which can be designed in various ways; · a voluntary third pillar, which is flexible and discretionary (this pillar, too, can take a variety of forms); and · a fourth pillar of informal intrafamily or intergenerational sources of financial and nonfinancial support to the elderly, including access to health care and housing. Pillar Design The reformed Polish pension system provides old-age income support to the elderly through all five of these pillars (table 6.2). The publicly man- aged, noncontributory zero pillar (social assistance benefit), which is financed with general tax revenues, redistributes income to lower-income groups using means testing. Both the traditional, publicly managed, pay-as- you-go first pillar and the privately managed, fully funded second pillar are earnings-related schemes for which participation is mandatory. The first pillar is an NDC scheme financed by mandatory contributions (paid by employees and employers in equal shares). The total contribution rate is 19.52 percent of wages, with 12.22 percentage points going to the NDC scheme and 7.30 percentage points going to the funded second pillar. The second pillar is a privately managed, fully funded defined-contribution scheme. Benefits from both schemes are payable in the form of annuities, the values of which are computed on the basis of contributions; credited investment earnings (which, in the case of the NDC scheme, are notional and based on the rate of growth in covered wages); and life expectancy at retirement. Supplementing the benefits of the first and second pillars is a voluntary third-pillar defined-contribution scheme. The mandatory fourth pillar is financed by a combination of contributions and general tax rev- enues and provides health insurance to the general population, including the elderly (table 6.3). Contributions to the first-pillar scheme are exempt from taxes while benefits are taxed. The second pillar is subjected to classical 186 Table 6.2 Structure of Poland's Pension System Taxation Investment income/ Benefit capital Scheme type Coverage Type Function Financing Generic benefit indexation Contributions gains Benefits Zero pillar Universal Means tested Redistributive Tax revenues Difference Regular increases n.a. n.a. Exempt (public between based on social noncontributory)a minimum assistance threshold and legislation actual income First pillar (public, Mandatory Notional Insurance Percentage of Pensions from Indexed to Exempt n.a. Taxed earnings related) defined- individual conversion inflation plus contribution earnings notional capital 20 percent of accumulation the growth in into annuities wages Second pillar Mandatory Defined Insurance Percentage of Pensions from Annuity Exempt Exempt Taxed (private, contribution individual conversion increased with earnings related) earnings capital 90 percent of accumulation return from into annuities investment from annuity reservesb Third pillar Voluntary Defined Insurance Voluntary Pension from Depends on Taxedc Exemptd Exempt (private, contribution contributions capital options chosen voluntary) accumulation Fourth Pillar Mandatory n.a. Insurance Percentage of Specified health n.a. Exempt n.a. n.a (public individual service package health care) earnings plus tax revenues Sources: Chlon, Gora, and Rutkowski 1999; European Commission 2007a, 2007b; OECD 2001; and data provided by the Ministry of Labour and Social Affairs. n.a. = Not applicable. a. There is also a minimum pension guarantee under the old-age pension system. For pensioners who contributed for at least 25 years (men) and 20 years (women) whose total pension falls below a certain threshold, the difference is topped up from the state budget. The guaranteed minimum benefit is currently 636 zlotys per month. Minimum pensions are taxed according to the general personal income tax rules. b. This increase is scheduled according to the draft law submitted by the government to the Parliament in 2008. c. Employer contributions to the third pillar are deductible from the employer's taxable income. d. Employees are granted tax relief up to 150 percent of average wage, above which they must pay taxes for capital gains and retirement savings. 187 188 Adequacy of Retirement Income after Pension Reforms expenditure taxation (exempt-exempt-taxed), meaning that contributions are exempt from taxation and investment income is exempt from tax- ation while benefits are taxed (see box 1.1 in chapter 1). The voluntary third pillar is subjected to taxed-exempt-exempt taxation, meaning that contributions and investment income are partially taxed while benefits are not taxed. The noncontributory zero pillar and the fourth pillar are not taxed. Noncontributory scheme. Like many of its neighbors, Poland has a non- contributory social assistance program that provides financial support to households in which income falls below a minimum threshold. In 2006, the threshold was about 14 percent of the average wage. Eligibility for social assistance benefits is not dependent on age. The elderly (including those collecting pensions) are not treated specially. For qualifying house- holds, the scheme provides two types of benefits: permanent benefits, payable to individuals incapable of working because of age or disability; and temporary benefits, payable to individuals incapable of working because of long-term illnesses, temporary disability, unemployment, or ineligibility for benefits from other social protection programs. Permanent benefits are means tested and adjusted to ensure that the household receives the mini- mum income threshold. Since October 2006, the maximum permanent benefit has been 444 zlotys (Zl) per month (18 percent of the average wage, which was Zl 2,477 per month in 2006). Temporary benefits are also means tested and adjusted for household income, but the benefit is equal to only 50 percent of the difference between household income and the minimum income threshold. Earnings-related schemes. Rather than changing the parameters of the existing, defined-benefit, pension scheme, Poland closed the scheme to younger workers (those born after 1948) and enrolled them in a new first-pillar scheme based on NDCs. Their pensionable rights that accrued under the old system until 1998 were converted into initial capital, which was credited to their individual accounts.5 Under the new scheme, benefits are computed using a defined- contribution formula, but the scheme's underlying financing remains on a pay-as-you-go basis. Individual account balances and investment earnings are purely notional (that is, they are an administrative record of contributions and credited interest without any underlying funds). Total old-age contributions are evenly split between employers and employees at the rate of 19.52 percent levied on wages up to 2.5 times the national Table 6.3 Parameters of Earnings-Related Schemes in Poland before and after Reform Contribution Pension Scheme type Period Vesting period Contribution rate ceiling Benefit rate assessment base Retirement age First Prereform 20 years for 45 percent paid by No ceiling on Flat component Best 3 65 for men, 60 pillar (earnings women, 25 employer, not contributions; (24 percent of consecutive for women related) years for men divided into ceiling of 2.5 times reference years from the (with many specific risk national average wage) plus 1.3 last 12 years exemptions, categories wage on benefit percent for since 1993, actual level each year of increasing retirement age contribution gradually to is about 59 for and 0.07 best 10 men and 55 for percent for consecutive women) each year of years from noncontributiona last 20 years Postreform No minimum 27.97 percentb 2.5 times national Pension from Notional capital 65 for men, 60 period (old-age: 19.52 average wage capital accumulation for women required; women percentc; disability accumulation are eligible for and survivor: minimum 6 percent; guarantee sickness and after 20 maternity: 2.45 years, men percent) after 25 years Prereform n.a. n.a. n.a. n.a. n.a. n.a. 189 (continued) 190 Table 6.3 Parameters of Earnings-Related Schemes in Poland before and after Reform (Continued) Contribution Pension Scheme type Period Vesting period Contribution rate ceiling Benefit rate assessment base Retirement age Second Postreform None; women 7.3 percent 2.5 times Pension from Accumulated 65 for men, 60 for pillar are eligible for national average capital funds women (earnings related) minimum wage accumulation guarantee after 20 years, men after 25 years Sources: European Commission 2007a, 2007b; consultations with World Bank staff. n.a. = Not applicable. a. As specified in the law, periods of university education, mandatory army service, maternity and child care leave, and unemployment are recognized. b. Individuals participating only in the first pillar pay 19.52 percent to the first pillar for old-age pensions (split equally between employers and employees). Individuals participating in both the first and the second pillars pay 12.22 percent to the first pillar (9.76 percent paid by employers and 2.46 percent paid by employees) and 7.3 percent to the second pillar (paid entirely by employees). Employers also pay contributions for work injury. The rate varies by industry. c. In 1999, all wages were increased by the contribution rate paid by employees in order to guarantee the same level of net income. This means that the rates of contributions are calculated on different bases and are therefore not directly comparable. Poland 191 average wage; 12.22 percent of wages is credited to the NDC pillar. The rate of credited "interest" is the rate of growth in covered wages in the sys- tem. At retirement, an individual's account balance is used to compute the amount of his or her benefit as a function of life expectancy. Actual benefits are paid from the contributions of current workers. The advantage of NDCs is that they improve incentives (by more tightly linking contributions with benefits and by making lifetime bene- fits conditional on lifetime wages rather than on a few best years of wages, as had been the case) and are fiscally self-balancing (inasmuch as accrued rights grow in proportion to the tax base used to fund them). Given the rapid aging of the Polish population, this aspect of NDCs is particularly important. Younger workers are also enrolled in a mandatory second-pillar scheme of funded individual accounts for which the contribution rate is 7.3 percent of wages. Benefits from both schemes are paid in the form of lifetime annuities. A minimum pension guarantee is awarded to men with 25 years of service and to women with 20 years of service. Voluntary scheme. The voluntary, defined-contribution, third-pillar scheme includes both employee pension funds and individual retirement accounts (table 6.4). Employee pension funds are registered and super- vised by the Insurance and Pension Funds Supervisory Commission. The state encourages employee pension plans by deducting the 7 percent con- tributed to the voluntary pillar from the wage base for the payment of social security contributions to the first pillar. Employees can make their own contributions, which are deducted directly from their wages and transferred to their accounts. Benefit eligibility begins at age 60. Benefits can be claimed as a lump sum or as a lifetime benefit. Employee pension plans are based on agreements with investment funds, insurance companies, or individual pension funds. As of April 2008, about 1,179 plans had been registered with the Financial Supervision Commission, 1,040 of which were active. Plans can be sponsored by one employer or by multiple employers with a minimum of five employees each. Employees must be 18 years old and have worked for the employer for at least three months to be eligible to participate. Employees with more than one employer may participate in multiple plans simultaneously. Individual retirement accounts were established in 2005. They can be held with open-ended investments funds, brokerage houses, banks, and insurance companies. Participants must contribute for at least five years 192 Adequacy of Retirement Income after Pension Reforms Table 6.4 Characteristics of the Voluntary Scheme in Poland Tax Lump-sum advantages Contributions payments Vesting Retirement to tax deductible possible in Coverage period age participantsa by employersa retirement Workers age 18 and older 5 Years 60 Yes Yes Yes Sources: European Commission 2007a; Leonik 2006. a. Participants are provided with tax relief up to 150 percent of average wage, above which they have to pay taxes for capital gains on retirement savings. As of January 2009, the ceiling will be increased to 300 percent of average wage. Employers are exempt from paying social security contributions to the first pillar for the 7 percent contribu- tion they pay to the voluntary pillar. and cannot withdraw benefits until age 60 (the same age applied to employee pension plans). Participants are provided with tax relief up to 150 percent of the national average wage, above which they pay taxes on capital gains on their retirement savings. By the end of 2007, more than 915,000 individuals (about 4 percent of the working-age population in Poland) had individual retirement accounts. Health care system. Health care in Poland is provided primarily through a mandatory health insurance system administered by the National Health Fund. Services are financed primarily by contributions from the covered population, at the rate of 9 percent of income, the definition for which varies by group.6 People receiving social insurance benefits--which include pensioners and recipients of social welfare allowances--pay con- tributions based on their gross benefits. Recipients of health care are required to make copayments for some health care services, pharmaceuticals, and medical devices (Kuszewski and Gericke 2005). In 2005, health expenditure accounted for 6.2 percent of GDP, 69.3 percent of which was public expenditure and 30.7 percent was private expenditure. Of the private expenditure, 85.1 percent was attrib- utable to out-of-pocket expenditure (informal payments, direct payments, and copayments) (WHO 2008). Institutional Structure and Coverage of Earnings-Related Schemes The mandatory first- and second-pillar schemes cover all salaried employ- ees and the self-employed (roughly 87 percent of the labor force in 2005). There are special systems for farmers, the police, and members of the military. The farmers' pension fund is administered by the Agricultural Social Insurance Fund. The Social Insurance Institution (ZUS) administers the first-pillar scheme through regional branches and local offices located Poland 193 throughout the country. The second pillar is administered by private pension-fund management companies. Contributions for both pillars are collected by ZUS, which transfers second-pillar contributions to the man- agement company of the participant's choice within five days. The Financial Supervision Commission is responsible for licensing and supervising open pension funds and pension-fund management compa- nies. In 2007, there were 15 open pension funds and pension-fund man- agement companies operating. The industry is concentrated, with the three largest companies controlling 64 percent of the market. Open pen- sion funds have become some of the largest institutional investors in Poland, with total assets equivalent to 12 percent of GDP at the end of 2007. This percentage compares favorably with other countries that have introduced multipillar reforms (Rocha and Rudolph 2007). Structure of Benefits Poland's pension system provides old-age, disability, and survivorship pensions. The provisions governing each of these types of benefits are discussed as follows. Old-age benefits. There is no work experience requirement to become eligible for an old-age pension; reaching the retirement age is sufficient. However, to become eligible for the minimum old-age pension guarantee from the mandatory system, men must have at least 25 years of service (women require 20 years) and have reached the retirement age of 65 (women can retire at age 60). Early retirement is not allowed. Delayed retirement is allowed without restriction. Workers who defer their pen- sion may continue to contribute to their notional and open pension-fund accounts in order to increase the amount of their pension. Upon a person's retirement, both first- and second-pillar accounts are converted into pensions. The time of conversion is the same for both pil- lars. Under the first-pillar NDC scheme, an individual's account balance is simply an administrative record of contributions and notionally cred- ited interest (based on the growth of economywide covered wages). At retirement, the account balance is used as the basis for computing the value of the individual's annuity based on life expectancy. Benefits are paid using the contributions of current workers. Benefits from the second pillar are also paid in the form of annuities but are funded using the accu- mulated capital in the individual's investment account. Annuities are cal- culated using unisex mortality tables and are increased by 90 percent of the interest earned by annuity companies on their annuity reserves, according to the draft law submitted by the government to Parliament in 194 Adequacy of Retirement Income after Pension Reforms June 2008. Men who have contributed for 25 years and women who have contributed for 20 years are eligible for a minimum pension in cases where their combined first- and second-pillar benefits would otherwise fall below the minimum pension (equal to 23 percent of the average wage in 2004).7 Minimum pensions are taxed according to the general personal income tax rules. Disability benefits. Disability benefits are determined on the basis of an individual's inability to work. They are provided to workers with at least 5 years of service credit over the 10 years before becoming disabled (for workers below age 30, the requirement is 1­4 years, depending on age), subject to the additional restriction that the noncontributory periods do not exceed one-third of the years of total contribution (table 6.5). Benefits for people who are totally disabled are calculated using the same rules used to compute old-age pensions. Benefits for people who are partially disabled are 75 percent of the amount awarded for total disability. Benefits are converted into an old-age pension upon reaching retirement. In 2008, the government proposed significant changes to the benefit formula to link disability pensions to old-age pensions based on the NDC Table 6.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in Poland under the First-Pillar Earnings-Related Scheme Contribution Contribution Partial Vesting period rate ceiling Eligibility Benefit rate pension Under age 20: 1 year 6 percent 250 percent Total or Flat component 75 percent Age 20­22: 2 years (4.5 percent of average partial (24 percent of total Age 22­25: 3 years by employer, wage incapacity of reference disability Age 25­30: 4 years 1.5 percent to workb wage) plus pension Over age 30: 5 years by employee)a 1.3 percent for each year of contributions and 0.07 percent for each noncontribu- tory year or years required to top up the total to 25 years Source: European Commission 2007a. a. This rate also covers survivor pension contributions. In 1999, the contribution rate for disability and survivor benefits was set at 13 percent, split equally between employees and employers. This was reduced in 2007 and 2008 to reach the current 6 percent level. b. Not defined in percentages. Poland 195 formula, with a view to making the system consistent across all of its component programs. The new formula will apply to pensions awarded from 2009 onward. Survivor benefits. Survivor benefits are awarded to dependents if the deceased had been receiving (or had met the criteria to receive) an old- age or disability pension. (table 6.6). Survivors who are contributing to the second pillar are entitled to the accumulated capital in the deceased's account, with 50 percent of the account balance going to the survivor's spouse and the remainder going to one or two other people named in the participant's contract with an open pension fund. For participants who do not name a beneficiary, the remainder is divided among the deceased's closest relatives (spouses and orphans are primary beneficiaries, parents and grandchildren are second- ary beneficiaries). The accumulated capital can be distributed as a lump sum or as installments over a two-year period, as specified by the benefi- ciary. For survivors of workers who had already been receiving an annu- ity from the second pillar, benefit eligibility (and the amount of the benefit) depends on the type of annuity that was purchased. Assessment of the Performance of Poland's Pension System The World Bank has established four principles for evaluating public pen- sion systems, which together should guide the process of pension reform (see Holzmann and Hinz 2005). Broadly speaking, these principles include the adequacy and security of benefits, the affordability of contributions, the sustainability of the system over time, and the robustness of the sys- tem in the face of demographic changes and macroeconomic shocks. This chapter focuses primarily on the adequacy of benefits and the financial sustainability of the earnings-related pension scheme. The remaining prin- ciples are mentioned only briefly. Adequacy is analyzed through the lens of net replacement rates. Financial sustainability is evaluated using projec- tions of pension expenditure and revenues. Benefit Adequacy Replacement rates are a useful yardstick for measuring the adequacy of pension benefits, because they express benefits relative to preretirement earnings, thereby indicating the degree to which income is replaced when workers retire. Two variants are commonly used. Gross replace- ment rates compute income replacement as the ratio of benefits paid to 196 Table 6.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in Poland under the First-Pillar Earnings-Related Scheme Spouse Orphan age Orphan Total family Eligibility replacement rate Benefit duration Remarriage test limit replacement rate benefit Eligibility of 85 percent of For life, if spouse is Benefits paid even 18 (25 for 85 percent of 90 percent of the deceased deceased's disabled, is taking if surviving university deceased's deceased's pension for an old-age pension if sole care of a child (until spouse remarries students) pension if sole for two survivors, or a disability surviving relative the child finishes surviving relative 95 percent pension school), or is above regardless of age 50; otherwise, number of one year survivors Sources: European Commission 2007a, 2007b. Note: An earnings test is conducted for survivor pensions. The amount of the benefit is reduced if a survivor is younger than the statutory retirement age or has an income of 70­130 percent of average national earnings. The benefit is suspended if the survivor's income exceeds 130 percent of average national monthly earnings. No income test is conducted once a survivor reaches the statutory retirement age (U.S. Social Security Administration 2006). Poland 197 pretax preretirement earnings. Net replacement rates compute income replacement as the ratio of benefits received (that is, after the pay- ment of taxes and other levies, including contributions for social insur- ance) to posttax preretirement earnings. In general, net replacement rates are a more useful measure of benefit adequacy, because they cap- ture the degree to which actual take-home pay is replaced when work- ers retire. The level of income replacement at retirement is not the only measure of benefit adequacy. For a full assessment of benefit adequacy, it is also important to determine how postretirement indexation rules will affect replacement rates during retirement. Pension benefits in retirement are expected to be indexed to inflation, so that their real value is maintained. In a growing economy with rising real wages, however, mere price index- ation of pensions leads to a deterioration of the relative consumption position of the retirees. For this reason, some countries have introduced mixed indexation of pensions that use varying weights of inflation and wage growth in the indexation formula. For an evaluation of the effect of indexation on replacement rates in Poland, the replacement rates are normalized to 100 percent and the assumptions for calculating the replacement rates are maintained (that is, inflation is 2.5 percent a year and real wage growth is 2.0 percent a year). The change in the replacement rate is measured in comparison with full wage indexation or the earnings of an active worker. The results of this analysis indicate that the relative income position of a retiree would deteriorate by 16 percent after 10 years in retirement and by 45 percent after 35 years in retirement. The evaluation of income replacement that follows considers replacement rates only at retire- ment; it does not take into account the impact of indexation policies on replacement rates during retirement. Replacement rates are a function of the formula governing pension benefits; an individual's contribution history; and, in the case of net replacement rates, the rules of income tax, social security contributions, and other relevant levies. The benefit formula establishes the degree to which the system redistributes income across individuals of different lev- els of preretirement earnings. Progressive systems provide higher levels of income replacement to people with lower levels of preretirement income. In general, the degree to which a system is redistributive depends on the existence (and value) of flat transfers and minimum pension guarantees, the degree to which benefits are earnings related, and the existence of ceilings on earnings subject to contributions. An individual's 198 Adequacy of Retirement Income after Pension Reforms contribution history can be characterized by his or her age of entry into the labor force, contribution density, and decisions regarding the timing of retirement. To some degree, these three factors are influenced by the incentives embodied in the pension system. The tax and contribution system influences net replacement rates through the progressiveness of the income tax formula, which taxes (higher) income during a worker's active life more so than it does (lower) pension benefits in retirement. In addition, there are are social security levies (for pensions; unemploy- ment; health care; and, at times, housing and family benefits), which are typically reduced or eliminated altogether in retirement. These benefits are particularly important for low- to middle-income groups. Benchmarks need to be established for an evaluation of the adequacy of the income replacement provided by the earnings-related pension schemes. Unfortunately, there is no consensus on what constitutes ade- quacy. According to one widely respected definition, pensions are adequate when they are sufficient to prevent poverty among the elderly and provide the vast majority of the population with a reliable mechanism for smooth- ing income over their lifetime. Even with a definition, however, establish- ing benchmarks is problematic, because attitudes vary across countries as a result of social and cultural perceptions. Moreover, benchmarks ignore the other factors that affect the welfare of the elderly--and that also vary across countries--including the existence and generosity of health insurance and long-term care, the cost of housing, the structure of traditional living arrangements, the presence of informal intrafamily or intergenerational sources of financial and nonfinancial support, and the availability and secu- rity of other mechanisms for people to save for their own retirement. One reputable nine-country study (OECD 2001) observes that liv- ing standards are roughly comparable for people 10 years older than the normal retirement age and people 15 years younger than the normal retirement age when retirees have disposable income equal to roughly 80 percent of the disposable income of working-age people. In part, this is attributable to the fact that retirees have no work-related expenses (they do not have to commute or buy special clothing or uniforms, for example). This finding, however, does not imply that mandatory first-pillar pension schemes should actually target an 80 percent net replacement rate. To the contrary, in middle- and high-income countries, one can reasonably expect individuals to save for their own retirement--and the empirical evidence suggests that, in practice, they do so.8 There is also some evidence to sug- gest that the ratio between preretirement and postretirement income is Poland 199 somewhat independent of the income replacement mandate of the public pension system. Put simply, individuals tend to save more in countries with more modest mandates (and vice versa). Because Poland has access to relatively well-developed financial mar- kets, it would seem reasonable to expect middle- and higher-income workers to save enough to finance at least 25 percent, if not closer to 50 percent, of this 80 percent income replacement target. Given this, three benchmarks are provided: a 40 percent net replacement rate (which implies that individuals would be expected to save enough to finance half of the total income replacement target); a 60 percent net replacement rate (which implies that individuals would be expected to finance a quarter of the target); and an 80 percent net replacement rate (which implies that individuals, most of whom would be low-income earners, would not be expected to contribute anything toward the target).9 In the following analysis, these three benchmarks are used to evaluate the adequacy of ben- efits in Poland compared with the average net replacement rate observed in 53 countries around the world, the average net replacement rate observed for selected countries in Europe and Central Asia, and the poverty line in Poland. To estimate gross and net replacement rates, we use the Analysis of Pension Entitlements across Countries (APEX) model to consider two critical dimensions: earnings levels and contribution periods.10 This model generates estimates for replacement rates under steady-state assumptions (that is, as if the rules of the reformed pension scheme had been in place over the entire active life of the individual). Because life expectancies at retirement are projected to increase over time--which will affect the benefits paid by defined-contribution pension schemes--a reference year must be chosen. The year 2040 is used, because it provides a sufficiently long contribution period to approximate steady-state conditions. The first critical task is to investigate levels of income replacement across a relevant spectrum of income. Income is represented as a per- centage (50­200 percent) of average earnings. The second task is to investigate the impact on income replacement of differences in the dura- tion, timing, and density of an individual's contribution history (density refers to the percentage of time an individual actually contributes over a given period). To facilitate the presentation of these multidimensional results, we compute replacement rates as a function of the age at which an individual exits the labor market. They are presented separately for full-career and partial-career workers. 200 Adequacy of Retirement Income after Pension Reforms Replacement rates for full-career workers. Full-career workers are examined first. For the purpose of this analysis, a full career is defined as continuous employment from age 20 to the current normal retirement age of 65 for men. Two earnings-related pension schemes contribute to gross replacement rates (figure 6.3).11 The two schemes directly connect the benefits an individual receives in retirement to the contributions he or she made while working. Regardless of income, gross replacement rates are 62.1 percent, 30.0 percentage points of which comes from the first pillar and 31.2 percentage points of which comes from the second. The situation does not change significantly when taxes are taken into consid- eration (figure 6.4). Pensions for most full-career workers in Poland can be considered adequate (figure 6.5). Replacement rates for workers of all levels of preretirement income are substantially higher than the middle bench- mark, which implies that the pension system is effectively smoothing consumption from work to retirement, especially for middle- and high-income workers.12 Replacement rates for low-income workers are 5 percentage points lower than the 80 percent benchmark. Given that benefits for even the lowest-income full-career workers greatly exceed the poverty line, the objective of poverty alleviation is being Figure 6.3 Sources of Gross Replacement Rates in Poland, by Income Level 80 70 60 (percent) 50 rate 40 30 replacement 20 10 gross 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (NDC) Source: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. Poland 201 Figure 6.4 Sources of Net Replacement Rates in Poland, by Income Level 100 90 80 70 (percent) 60 rate 50 40 30 20 replacement 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (NDC) taxes Source: APEX model. Note: Figure shows projected replacement rate for full-career worker in 2040 as approximation of steady-state conditions. met.13 Levels of income replacement are higher than the world and regional averages for high-income workers but lower than both averages for low-income workers. This finding indicates that the strong link between contributions and benefits found in the Polish pension system also results in the system affecting comparatively little redistribution from workers with high preretirement income to those with lower pre- retirement income. Replacement rates for partial-career workers. Not everyone works from age 20 to the statutory retirement age. Many individuals enter and exit the labor force (often at different ages and for different periods of time) and earn different wages while working (figure 6.6). To examine the ade- quacy of benefits for partial-career workers, we consider three stylized cases. These cases include career type A (someone entering the labor force at age 25 who works continuously for a period of years before leaving the workforce at some point between the ages of 50 and 70 and then claims a benefit); career type B (identical to career type A, except that the worker enters the workforce at age 30 and leaves no earlier than age 55); and career type C (identical to career type A, except that the individual contributes in only three years out of four while in the labor force). For 202 Adequacy of Retirement Income after Pension Reforms Figure 6.5 Net Replacement Rates for Male Full-Career Workers in Poland, Europe and Central Asia, and the World, by Income Level 100 high benchmark 80 (percent) 60 middle benchmark rate 40 low benchmark replacement 20 net 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage world average Europe and Central Asia average Poland average poverty line (percentage of average income) Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. cases in which the withdrawal from the formal labor market occurs before the statutory retirement age, the pension is claimed (and the replacement rate calculated) only at the later age. For withdrawals after the statutory retirement age, the ages coincide. Four conclusions can be drawn from examination of net replacement rates for partial-career middle-income workers (only middle-income partial-career workers are examined because replacement rates are com- parable for workers with lower and higher levels of preretirement income). First, all three types of workers receive levels of income replacement that exceed the poverty line. Second, leaving the workforce very early can be very costly. Third, entering the workforce later in life is costly, because a worker who enters the workforce at age 30 receives a net replacement rate 9­12 percentage points lower than a worker who enters at age 25. Fourth, working intermittently is costly, because contributing three years out of every four results in a net replacement rate that is 9­23 percentage points lower than when contributing continuously. In all cases, net replacement rates grow faster the longer an individual works. This is encouraging, because it provides incentives for individuals to Poland 203 Figure 6.6 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Poland, by Career Type and Exit Age 110 high benchmark 100 world and Europe and Central Asia average 90 80 (percent) 70 60 rate 50 40 middle benchmark 30 low benchmark replacement 20 net poverty line (percentage of average income) 10 0 45 50 55 60 65 70 exit age from labor market career type A: career type B: career type C: entry age--25, entry age--30, entry age--25, contribution contribution contribution density--100% density--100% density--75% Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. See text for descriptions of career types. defer retirement. Although career type A workers can attain the 60 percent benchmark before reaching the normal retirement age, career type B workers must work one year beyond the normal retirement age and career type C workers must work four additional years to reach this benchmark. To replace 80 percent of preretirement earnings, career type A workers must work for up to three years past the normal retirement age, while career type B workers must work until age 70. Career type C work- ers cannot attain the 80 percent benchmark even if they work until age 70. Fiscal Sustainability The sustainability of a pay-as-you-go first-pillar pension scheme is best evaluated in actuarial terms by estimating the scheme's actuarial deficit as the difference between its assets and liabilities. If a large actuarial deficit exists, the scheme is financially unsustainable and needs policy actions that increase its assets, reduce its liabilities, or both. A good proxy for the actuarial deficit is the difference between the present value of the scheme's expected future revenues (that is, contributions and other 204 Adequacy of Retirement Income after Pension Reforms Figure 6.7 Projected Fiscal Balance of Poland's Public Pension Scheme, 2004­50 20 15 GDP 10 of 5 0 percentage ­5 ­10 2004 2010 2015 2020 2030 2040 2050 year expenditures revenues balance Source: European Commission 2007b. Note: Projections 1998 and cover old-age pensions only. income) and expected future expenditures (that is, benefit payments, administrative costs, and other expenses) over an extended projection period. The difference between these two values represents an unfunded liability (sometimes referred to as a financing gap) on the public-sector balance sheet. Because this study is also concerned with the time path of revenues and expenditures (and the resulting balance across the projec- tion period ending in 2050), this more pragmatic approach has been taken. Projections of expenditures, revenues, and deficits are presented on the basis of available postreform fiscal projections. The reforms of 1999 eliminated the pension system's unfunded liabil- ity. As a result, the system will eventually come very close to fiscal bal- ance (figure 6.7). Over time, projected revenues will increase slightly, to about 8.8 percent of GDP by 2050. Expenditures are expected to fall more substantially, eventually reaching an amount equivalent to 9.6 per- cent of GDP, down from about 14.0 percent of GDP in 2004. As a result, the pension system is expected to generate a deficit of about 0.8 percent of GDP in the long term, down from about 6.1 percent in 2004. Conclusions To address lingering deficits in the pension system and projections that show that the aging of the population will put increasing pressure on the finances of the system over time, Poland introduced a multipillar pension Poland 205 system in 1999. The existing traditional pay-as-you-go public pension scheme was replaced with a mandatory, NDC first-pillar scheme; a mandatory, privately managed, defined-contribution second-pillar scheme; and a voluntary, fully funded, defined-contribution third-pillar scheme, intended to provide workers with a mechanism for saving outside of the mandated schemes. The mandated schemes directly connect benefits received in retirement to an individual's contributions made while work- ing, thereby improving incentives to participate and to remain in the workforce (albeit at the cost of affecting comparatively little redistribu- tion from high-income workers to lower-income workers). Poland's approach to pension reform was more radical than the approaches taken by its neighbors, many of which elected only to change the parameters of their existing defined-benefit schemes. Poland's pro- gram of reform was named "Security through Diversity," because its mul- tipillar approach effectively diversified retirement savings by enabling benefits to be paid both from a pay-as-you-go scheme (which, over time, will become much smaller than it is currently) and from a funded scheme (the returns of which are not perfectly correlated). Together, Poland's reforms improved the fiscal position of the pension system. The improvement in the fiscal balance of the pension system results from higher standard retirement ages than are found in most countries in the region (65 for men and 60 for women) and from automatic reduc- tions in replacement rates in step with increases in life expectancy. This incentive structure of both the first- and the second-pillar schemes should motivate individuals to postpone their retirement in line with increases in their life expectancy at retirement and should increase labor force partic- ipation across all ages (see Chlon-Dominczak 2009). Enabling elderly workers to continue to participate in the labor market, however, will require the introduction of cross-sectoral policy reforms.14 These improvements in the fiscal balance of the scheme notwith- standing, the resulting gross and net replacement rates for full-career workers are fully in line with regional and world averages (see chapter 1). The projected net replacement rates are close to the high benchmark of 80 percent. As in other countries, workers with less than full careers-- because they leave the workforce before reaching retirement age, work intermittently, or have gaps in their employment history--risk receiving a level of income replacement that is closer to--or even below--the lower benchmark of 40 percent. All workers have the option of saving outside of the two mandated schemes by participating in the voluntary third-pillar pension scheme. To increase income replacement by 1 percentage point, full-career workers 206 Adequacy of Retirement Income after Pension Reforms would need to save only about 0.5 percent of their wages between age 40 and the retirement age. Given that participation in the third-pillar scheme remains low, an opportunity exists for broadening its reach. Notes 1. The reform was named "Security through Diversity," because the multipillar approach effectively diversified retirement savings by enabling benefits to be paid from both a pay-as-you-go pillar and a funded pillar, the returns of which are not perfectly correlated (see Chlon, Gora, and Rutkowski 1999). 2. An NDC pension scheme is financed on a pay-as-you-go basis, but its bene- fits are computed using a defined-contribution formula. Under an NDC scheme, an individual's account "balance" and investment "earnings" are purely notional (that is, an administrative record of contributions and credited interest without any underlying funds). The rate of credited "interest" is often based on some economic proxy, such as the rate of growth in average wages. At retirement, an individual's account balance is used to compute the amount of his or her benefit, as a function of life expectancy. Actual benefits are paid from the contributions of current workers, as in all pay-as-you-go pension schemes. The advantage of NDCs is that they improve incentives (by more tightly linking contributions with benefits) and, to some degree, are fiscally self-balancing, inasmuch as accrued rights grow more in line with the resources required to fund them (see Holzmann and Palmer 2006). 3. Most occupations permitted workers to retire up to five years before reaching the legal retirement age. Some offered even more liberal provisions. Miners, for example, could retire after 25 years of service and teachers after 30 years, regardless of age; ballet dancers could retire at age 38 (see Chlon, Gora, and Rutkowski 1999). 4. For more information regarding these population projections, see Reiterer (2008). 5. Individuals covered by the new system who accrued rights under the old sys- tem by the end of 2008 can retire under the old rules. The 2008 cutoff was initially proposed to end in 2006; on the basis of decisions by the Parliament in 2005 and 2007, it was extended for two more years to address the lack of regulations related to the conversion of some early retirement options into so-called "bridging pensions." 6. Salaried employees make contributions based on their total income (that is, there is no ceiling on income for contribution purposes). The self-employed make contributions on the amount of their declared income or on 75 percent of the average wage, whichever is higher. The basis for calculating health insurance contributions for farmers is the price of 0.5 quintals of rye per stan- dard hectare of the farm. Poland 207 7. The difference between the minimum pension and the benefit the individual would have received is financed from general tax revenues. The minimum pension is indexed according to general indexation rules. 8. In Chile, for instance, 70 percent of retirees from the mandatory public pen- sion system own their home, which is a form of savings (Valdés-Prieto 2008). 9. These benchmarks approximate the standards developed by the International Labour Organization (ILO) (1952) and the Council of Europe (1990). ILO Convention 102 of 1952 sets a minimum benefit equal to 40 percent of the reference wage for married men of pensionable age. This amount was raised to 45 percent in 1968. The European Code of Security of 1990 sets a mini- mum standard for members of the Council of Europe equal to 65 percent for married people of a specific age. 10. The APEX model was developed by Axia Economics, with funding from the Organisation for Economic Co-operation and Development and the World Bank. The model codes detailed eligibility and benefit rules for first- and second-pillar schemes based on available public information that has been verified by country contacts. Because the details of the rules sometimes change on short notice (and limited public disclosure), the calculations presented here should be considered as best approximations only. 11. Replacement rates are simulated for an unmarried man working a hypothet- ical career path under the assumption that real wage growth is 2 percent, inflation is 2.5 percent, the rate of return on invested assets is 3.5 percent, and the worker retires at the statutory retirement age. 12. Estimates are based on the assumption that real wage growth is 2 percent, the net real rate of return on invested assets is 3.5 percent, and benefits (from the unfunded and funded pillars) are price indexed. 13. As a proxy for the poverty line, a figure of 35 percent of the average net wage is used, because this percentage broadly approximates a US$2.25-a-day poverty line converted into national currency, adjusted for purchasing power parity, expressed relative to the national average net wage, and averaged across the eight study countries. Such an approach enables valid comparisons to be made across the sample (see chapter 1). 14. See Holzmann, MacKellar, and Repansek (2009) for a conference volume that addresses theses issues for the countries of southeastern Europe. Bibliography Chlon A., M. Gora, and M. Rutkowski. 1999. Shaping Pension Reform in Poland: Security through Diversity. Washington, DC: World Bank. Chlon-Dominczak, A. 2009. "Pension System and Employment of Older Workers: How to Change the Incentive Structure? The Polish Experience." In Pension 208 Adequacy of Retirement Income after Pension Reforms Reform in Southeastern Europe: Linking to Labor and Financial Market Reforms, ed. R. Holzmann, L. MacKellar, and J. Repansek, 163­76. Washington, DC: World Bank. Council of Europe. 1990. European Code of Social Security (Revised). Rome. European Commission. 2006. "The Impact of Ageing on Public Expenditure: Projections for the EU-25 Member States on Pensions, Health Care, Long- Term Care, Education, and Unemployment Transfers (2004­2050)." In European Economy Special Report 1, Economic Policy Committee, Brussels. ------. 2007a. Mutual Information System and Social Protection (MISSOC) database. http://ec.europa.eu/employment_social/. ------. 2007b. "Pension Schemes and Projection Models in EU-25 Member Countries." European Economy Occasional Paper 37, Economic Policy Committee and Directorate General for Economic and Financial Affairs, Brussels. GVG (Gesellschaft für Versicherungswissenschaft und -gestaltung) and European Commission. 2003. Study on the Social Protection Systems in the 13 Applicant Countries: Poland Country Report. Cologne. Gora, M., and M. Rutkowski. 2000. The Quest for Pension Reform: Poland's Security through Diversity. Working Paper No. 286, University of Michigan. Holzmann, R., and R. Hinz. 2005. Old-Age Income Support in the 21st Century. Washington, DC: World Bank. Holzmann, R., and E. Palmer, eds. 2006. Pension Reform: Issues and Prospects for Non-Financial Defined Contribution Schemes. Washington, DC: World Bank. Holzmann, R., L. MacKellar, and J. Repansek, ed. 2009. Pension Reform in South- eastern Europe: Linking to Labor and Financial Market Reforms. Washington, DC: Word Bank. INPRS (International Network of Pension Regulators and Supervisors). 2003. Complementary and Private Pensions throughout the World. Geneva: INPRS. ILO (International Labour Organization). 1952. ILO Convention 102. Geneva: ILO. ------. 1967. ILO Convention 128. Geneva: ILO. Kuszewski, K., and C. Gericke. 2005. Health Care Systems in Transition: Poland. World Health Organization Regional Office for Europe, Copenhagen. Leonik, A. 2006. "Pension Security System in Poland." http://www.seminar- burzy.cz/files/leonik03-06.pdf. OECD (Organisation for Economic Co-operation and Development). 2001. Ageing and Income: Financial Resources and Retirement in 9 OECD Countries. Paris: OECD. Putelbergier, B. 2000. "Changes in the Financing of the Pension System in the Face of Aging: The New Pension System in Poland and Its Adaptation to Other Poland 209 Countries' Systems." Paper presented at the International Social Security Association Conference on Social Security, Helsinki, September 25­27. Reiterer, A. 2008. Population Development and Age Structure in Southeastern Europe until 2050. World Bank, Washington, DC. Rocha, R., and H. Rudolph. 2007. "Competition and Performance in the Polish Second Pillar." Working Paper 107, World Bank, Washington, DC. ZUS (Social Insurance Institution). 2006. Social Insurance in Poland: Information, Facts. Warsaw. U.S. Social Security Administration. 2006 Social Security Systems throughout the World: Europe. Washington, DC: Social Security Administration. Valdés-Prieto, S. 2008. Designs for the First Pillar Pensions and the 2008 Chilean Reform. http://editorialexpress.com/cgi-bin/conference/download.cgi?db_ name=SECHI2008&paper_id=130. Whitehouse, E. 1999. Tax Treatment of Funded Pensions. Washington, DC: World Bank. World Bank. 2005. Growth, Poverty and Inequality: Eastern Europe and the Former Soviet Union. Washington, DC: World Bank. ------. 2007a. Pensions Panorama. Washington, DC: World Bank. ------. 2007b. Social Assistance in Central Europe and the Baltic States. Washington, DC: World Bank. WHO (World Health Organization). 2008. Database. http://www.who.int/ research/en/. C H A P T E R 7 Romania Romania inherited a socialist-era public pension system financed on a pay-as-you-go basis (meaning that contributions from current workers are used to pay benefits to current beneficiaries). As a result of increased informality in labor markets following its transition from a centrally planned economy to a market economy, pension system revenues fell from the equivalent of 9.3 percent of gross domestic product (GDP) in 1992 to 6.3 percent of GDP by 1998--despite an increase in contribu- tion rates--while expenditure fluctuated at about 7 percent of GDP. As a result, net cash flow fell from a surplus of 1.9 percent of GDP to a deficit of 0.8 percent of GDP. Recognizing that a pension system that offered comparatively generous benefits, low retirement ages, and lax benefit eligibility conditions would not be sustainable over the long term as the population ages, the govern- ment introduced substantial reforms to the pension system in 2000. These reforms raised retirement ages, extended the service period required to become eligible for a full pension, imposed new conditions on early retire- ment, and replaced the traditional defined-benefit formula with a new for- mula based on points. In 2006, the government passed legislation to introduce a mandatory, fully funded, defined-contribution scheme, which became operational in 2008. 211 212 Adequacy of Retirement Income after Pension Reforms Against this backdrop, this chapter evaluates Romania's pension sys- tem, focusing on fiscal sustainability and benefit adequacy. Adequacy is evaluated through the lens of statutory net replacement rates for differ- ent retirement ages, patterns of contributions, and income levels relative to international benchmarks. This chapter is organized as follows. The next section discusses the motivation for the reforms. The following section describes the key char- acteristics of the reformed pension system. The third section assesses the adequacy of pension benefits and the fiscal sustainability of the system. The last section draws conclusions. Motivation for Reform The pension system that Romania inherited suffered from a number of serious design flaws similar to those observed in other transition economies. These included (a) low retirement ages, which allowed pen- sioners to receive benefits for very long periods; (b) generous benefits, which created incentives for early retirement, further increasing the period over which benefits were paid; (c) the computation of benefits based on an individual's last five years of wages rather than his or her lifetime wages, which weakened the link between contributions and benefits and created incentives for workers to underreport income or to migrate from the for- mal to informal sectors; and (d) lax eligibility conditions governing the award of disability pensions, which resulted in more beneficiaries than could be justified on the basis of impairment. The pension system was also fragmented and provided special privileges for particular occupations. In the early years of the transition to a market economy, pension sys- tem revenues in Romania fell, as a result of increasing informality in labor markets and the restructuring of state enterprises, which contributed to rising informal-sector employment. As a result, the number of individuals contributing to the pension system fell from about 8 million in 1990 to 5 million by 1999. To ease the impact of enterprise restructuring, the gov- ernment granted many workers early retirement. As a result, the number of beneficiaries rose from 2.2 million to 4.0 million over the same period. To cope with these changes, the government increased the contribution rate from 14 percent in 1990 to 28 percent in 1992 (World Bank 2004). Despite the change, revenues still fell short of expenditures, and by 1995, the scheme generated a deficit of 0.2 percent of GDP (table 7.1). The deficits of the pension system in the late 1990s were not huge and could probably have been afforded were it not for the fact that they were Romania 213 Table 7.1 Fiscal Balance of Romania's Pension System before Reform, 1992­98 (percentage of GDP) Year Revenues Expenditure Balance 1992 9.3 7.4 1.9 1993 8.1 6.7 1.4 1994 7.1 6.7 0.4 1995 6.8 7.0 ­0.2 1996 6.7 6.9 ­0.2 1997 6.5 6.5 ­0.0 1998 6.3 7.1 ­0.8 Source: World Bank 2004. expected to grow substantially over the medium and long terms as a result of the aging of the population.1 The aging of Romania's population is captured in the old-age dependency ratio (the population age 65 and older divided by the population age 20­64), which is projected to rise from 23.6 percent in 2005 to 56.6 percent by 2050.2 Recognizing these challenges, the government introduced substantial reforms to the pension system in 2000. These changes included raising retirement ages, extending service periods for eligibility for full pensions, imposing new conditions for early retirement, and replacing the tradi- tional defined-benefit formula with a new formula based on points. These measures managed--at least in the short run--to balance revenues and expenditures and to achieve a fragile surplus equivalent to 0.3 percent of GDP in 2006 and 0.2 percent in 2007. In 2006, the government passed legislation to introduce a mandatory fully funded defined-contribution scheme, which became operational in May 2008. Characteristics of Romania's Pension System This section describes the main characteristics of Romania's pension sys- tem. They include the design of the individual pillars of social insurance; the rules governing pension system taxation, institutional structure, and coverage; and the provisions governing old-age, disability, and survivor- ship pensions. The design of the pension system is assessed using a conceptual frame- work developed by the World Bank, which generally recommends includ- ing a funded component if conditions are appropriate but increasingly recognizes that a range of choices is available to policy makers to provide 214 Adequacy of Retirement Income after Pension Reforms effective old-age protection in a manner that is fiscally responsible (see Holzmann and Hinz 2005). In general, the World Bank supports pension systems composed of some combination of five basic pillars: · a noncontributory (or zero) pillar (in the form of a demogrant, social pension, or social assistance benefit) intended to provide a minimal level of income protection; · a first-pillar contributory system linked to earnings, which seeks to re- place a portion of preretirement income; · a mandatory second pillar (essentially, individual savings accounts), which can be designed in various ways; · a voluntary third pillar, which is flexible and discretionary (this pillar, too, can take a variety of forms); and · a fourth pillar of informal intrafamily or intergenerational sources of financial and nonfinancial support to the elderly, including access to health care and housing. Pillar Design The design of Romania's pension system incorporates all five of the pillars recommended by the World Bank (table 7.2). The publicly managed non- contributory zero pillar, financed with general tax revenues, redistributes income to lower-income groups using means testing that considers both income and assets. The amount of the noncontributory benefit is based on the state-defined minimum income guarantee, which is adjusted by the government on the basis of inflation. Both the publicly managed, pay-as-you-go first pillar and the newly introduced, privately managed, fully funded second pillar are earnings- related schemes. Benefits under the first pillar are calculated from an indi- vidual's accumulated points, which are determined by his or her wages relative to the average wage. Second-pillar benefits are a function of an individual's contributions and investment earnings; the procedures govern- ing the payout of benefits are yet to be established. Contributions for the second pillar started in May 2008. The third pillar is an optional privately managed, fully funded, defined-contribution pension scheme, which is intended to provide individuals with a mechanism for supplementing the benefits paid by the mandatory pillars. The fourth pillar provides health care to the elderly as part of the national health care system. First-pillar contributions are exempt from taxation while benefits are taxed. The fully funded second and third pillars will be subjected Table 7.2 Structure of Romania's Pension System Taxation Investment income/ Benefit capital Scheme type Coverage Type Function Financing Generic benefit indexation Contributions gains Benefits Zero pillar Universal Means tested Redistributive Tax revenues Difference between Government n.a. n.a. Exempt (public minimum income decision based noncontributory) guarantee and on changes in actual income consumer price index First pillar Mandatory Points Insurance Percentage Benefit calculated Adjusted on the Exempt n.a. Taxed (public, earnings of individual on the basis of the basis of changes related) earnings number of points in point valuea earned Second pillar Mandatory Defined Insurance Percentage Pension from capital Regulation Exempt Exempt Taxed (private, earnings contribution of individual accumulation on benefits does related) earnings not exist yet Third pillar (private, Voluntary Defined Insurance Voluntary Pension from capital Regulation Exemptb Exempt Taxed voluntary) contribution contributions accumulation on benefits does not exist yet Fourth pillar (public Mandatory n.a. Insurance Percentage Specified health n.a. Exempt n.a. n.a. health care) of individual service package earnings plus tax revenues Sources: European Commission 2007; OECD n.d. n.a. = Not applicable. 215 a. The point value cannot fall below 45 percent of the gross average wage. The percentage is adjusted based on ad hoc decisions by the government. b. An amount up to 200 euros per year per participant is tax exempt. 216 Adequacy of Retirement Income after Pension Reforms to exempt-exempt-taxed taxation (that is, classic expenditure tax), meaning that contributions are exempt from taxation (partially for the third pillar) and investment income is exempt but benefits are taxed (see box 1.1 in chapter 1). The zero and fourth pillars are completely tax exempt. Noncontributory scheme. Romania does not have a noncontributory social protection scheme specifically for the elderly, but the elderly are eligible for the minimum-income guarantee program, which provides financial support to households whose income falls below a minimum threshold. The threshold is a function of household size and income; the amount of the benefit is adjusted to make up the difference between the minimum income threshold and actual household income (World Bank 2003). In 2006, the monthly benefit was leu 92 (about 9 percent of the average wage) for a one-person household and leu 166 (15 percent of the average wage) for a two-person household. Benefits for large households rise in diminishing amounts. In 2005, 834,000 beneficiaries received minimum- income guarantee benefits, at a cost of leu 472 million, equivalent to 0.2 percent of GDP (World Bank 2007b). Earnings-related schemes. Both the publicly managed, pay-as-you-go first pillar and the newly introduced privately managed, fully funded second- pillar are earnings-related schemes (table 7.3). In 2000, the defined-benefit formula used to calculate pensions under the traditional first-pillar scheme was eliminated in favor of a new formula based on points. Under this formula, an individual's points are determined by his or her wages relative to the average wage. Benefits are based on total accumulated points at retirement.3 The number of years of wages on which benefits are based is gradually increasing, from the best five years to the entirety of an individual's service, thereby improving transparency and tightening the link between lifetime contributions and the benefits received in retire- ment. Under the law, the point value cannot fall below 45 percent of the gross average wage--the exact value of which is set by the government on an ad hoc basis. Pensions paid to existing beneficiaries are also adjusted on the basis of changes made to the point value.4 Retirement ages are gradually being raised from 62 to 65 for men and from 57 to 60 for women, a change that is being implemented so slowly that it will not be fully implemented until 2015. Moreover, although the new retirement age for men is consistent with international norms, the age for women remains low by international standards. By 2050, Romanian women will collect benefits for 50 percent longer than men.5 Table 7.3 Parameters of Earnings-Related Schemes in Romania before and after Reform Vesting Contribution Pension Retirement Scheme type Period period Contribution rate ceiling Benefit rate assessment base age First pillar Prereform 10 years 25.5 percent 5 times the 2.5 percent accrual rate for 5 best consecutive 62 for men, 57 (earnings (employee-employer average men for the first 30 years and years in last for women related) breakdown not wage 1 percent for each year there- 10 years available) after; 3 percent accrual rate for women for the first 25 years and 1 percent for each year thereafter Postreform 15 years 29.5 percenta None Benefit calculated on the basis Lifetime average 65 for men, 60 (20.0 percent by of the number of points indexed to for women employer, earned nominal wage 9.5 percent by growth employee) Second pillar Prereform n.a. n.a. n.a. n.a. n.a. n.a. (earnings Postreform Not yet 2 percent increasing None Pensions from capital Accumulated 65 for men, 60 related) established to 6 percent (over a accumulation funds for women by law period of 8 years) Sources: European Commission 2007; World Bank 1998, 2004. n.a. = Not applicable. a. The total contribution rate for old-age, disability, and survivor pensions is 29.75 percent. Individuals participating only in the first pillar pay 29.5 percent to the first pillar (20.0 percent by employer, 9.5 percent by employee). Individuals participating in both the first and the second pillars pay 27.5 percent to the first pillar (18.0 percent by employer, 9.5 percent by employee) and 2 percent rising to 6 percent to the second pillar (paid entirely by the employer). In 2009, the contribution rate will be reduced to 28 percent (18.5 percent by employer, 9.5 percent by employees). 217 218 Adequacy of Retirement Income after Pension Reforms Eligibility for a full pension now requires 35 years of service for men and 30 years for women, up from 30 years and 25 years, respectively, under the old system. The reform reduces costs and creates incentives for individuals to remain in the workforce longer (retiring with higher bene- fits). The minimum contribution period required to become eligible for benefits (the vesting period) was raised from 10 to 15 years, a change that will not be fully implemented until 2015. In 2006, the government passed legislation to introduce the fully funded, defined-contribution, second-pillar pension scheme. This scheme is mandatory for everyone up to age 35 and voluntary for people age 36­45. Contributions will be diverted from the first pillar, starting at 2 percent in 2008 and increasing by half a percentage point each year until the rate reaches 6 percent in 2016.6 Voluntary scheme. A voluntary, privately managed, fully funded third-pillar pension scheme was introduced as part of the government's reform program to provide individuals with a mechanism for supplementing the benefits paid by the mandatory pillars (table 7.4). Benefits payable under the third-pillar scheme will be a function of an individual's con- tributions and investment earnings at retirement. The mechanism by which benefits will be paid under the scheme will be discussed in 2009. Total contributions are limited to 15 percent of gross monthly salary (total of employer and employee). Contributions can be made by employees or employers on the basis of agreements between the parties or existing labor contracts. Contributions are deductible (for both employees and employers), up to 200 euros per year. Benefit eli- gibility requires that participants reach age 60, have contributed for at least 90 months, and have accumulated capital sufficient to meet a minimum threshold. In the event of disability before retirement, a Table 7.4 Characteristics of Romania's Voluntary Scheme Contributions Lump-sum Tax tax payments Vesting Retirement advantages to deductible by possible in Coverage period age participants employers retirement Employees and the 90 months 60 Yes Yes Regulation self-employed on benefits does not exist yet Source: OECD n.d. Romania 219 participant is entitled to receive the funds in his or her account. In the event that the participant dies before reaching retirement, account funds will be distributed to the participant's surviving dependents. Seven pension fund management companies are currently sponsor- ing voluntary pension funds. The largest three manage 73 percent of the assets in the scheme (Romania Private Pension Supervisory Commission Web site [http://w4.csspp.ro/en/]). At the end of 2007, 50,887 individu- als (0.5 percent of the labor force) were participating in the scheme, and assets totaled leu 14.3 million (less than half of one percent of GDP). The scheme appears to be growing quickly. By March 2008, the number of participants had increased to 75,423, and total assets had reached leu 24.8 million. Health care system. Health care in Romania is provided primarily through mandatory health insurance. Voluntary health insurance is available, but it is purchased mainly for travel abroad to countries in which services are not covered by Romania's mandatory scheme. The mandatory scheme is administered by district health insurance funds, which are responsible for collecting contributions and reimbursing claims from providers for health care services in their respective districts. The funds are regulated by the National Health Insurance Fund. The system is financed primarily by contributions from the covered population. The contribution rate for employed people is 14 percent of payroll, split equally between employers and employees. The contribu- tion rate for self-employed people, farmers, and pensioners is 7 percent. Children, people with disabilities, war veterans with no income, and the dependants of insured people do not pay for coverage. Recipients of health care services are required to make copayments for some medical services and pharmaceuticals (WHO 2000). In 2005, health expenditure accounted for 5.5 percent of GDP, 70.3 percent of which was public expenditure and 29.7 percent was private expenditure. Of the private expenditure, 85.0 percent was attributable to out-of-pocket expenditure (informal payments, direct payments, and copayments) (WHO 2008). Institutional Structure and Coverage of Earnings-Related Schemes The first-pillar pension scheme covers employees with individual labor contracts, civil servants, judges, cooperative members, and recipients of unemployment benefits. There are special schemes for some professions, including lawyers and members of the military. In 2005, 5.9 million 220 Adequacy of Retirement Income after Pension Reforms individuals (39.1 percent of the working-age population and 57.6 percent of the labor force) contributed to the scheme. The National House of Pensions is responsible for collecting contribu- tions and paying benefits. It will also be tasked with collecting second- pillar contributions and transferring them to the appropriate private pension fund management company. The Romanian Private Pension System Supervision Commission is responsible for licensing and regulat- ing the activities of the private pension companies. As of October 2007, six companies had applied for licenses to operate second-pillar schemes. Structure of Benefits The first-pillar earnings-related pension scheme provides old-age, disabil- ity, and survivorship pensions.7 The provisions governing each of these types of benefits are discussed below. Old-age benefits. Eligibility for a reduced old-age pension under the first- pillar scheme currently requires individuals to have at least 11 years and 2 months of contributory service. This requirement is gradually being increased to 15 years by 2015. Eligibility for a full pension requires 30 years and 9 months of service for men and 25 years and 9 months of service for women. This requirement is gradually being increased to 35 years for men and 30 years for women by 2015. The retirement age is 62 years and 9 months for men and 57 years and 9 months for women. Retirement ages are gradually being increased to 65 for men and 60 for women by 2015. Individuals may retire up to five years before reaching their retirement age, subject to a reduction in benefits, provided they have contributed for 10 years more than the number of years of contributory service required to earn a full pension (European Commission 2007). Old-age benefits are based on a point system. Points are awarded each year on the basis of an individual's wages divided by the average wage. At retirement, an individual's total accumulated points are divided by his or her total years of service. This value is then multi- plied by the pension point value to determine the individual's benefit. Under the law, the point value must not fall below 45 percent of the gross average wage, the exact value of which is set by the government on an ad hoc basis. Disability benefits. Disability pensions are awarded to individuals who have lost at least 50 percent of their capacity to work. Participants who achieved the contributory period identified are entitled to a disability Romania 221 pension (table 7.5). Participants entitled to a disability pension are granted a potential contributory period representing the difference between the full contributory period and the actual period of contribution at the time of disability. There are three categories of disability depending on the degree of incapacity. The first, second, and third categories are awarded 0.75, 0.60, and 0.40 points per year, respectively. Upon reaching retirement age, recipients of disability benefits can continue to receive their benefits or elect to receive an old-age pension instead. Under the second-pillar scheme, participants who become disabled will be entitled to a lump-sum payment or periodic payments for up to five years if their account is insuf- ficient for a minimum payment. Otherwise, participants can collect the pension they are entitled to from the second pillar. Survivor benefits. Survivor benefits are awarded to spouses and orphans of individuals who, at the time of their death, were receiving (or had met the criteria to receive) an old-age or disability pension (table 7.6). Upon reaching the retirement age, spouses are entitled to 50 percent of the deceased's pension if the spouse had been married for at least 15 years. Spouses married for 10­15 years are entitled to reduced benefits. Benefits are reduced by 0.5 percent a month for each month short of 15 years of marriage. Disabled spouses are entitled to survivor benefits regardless of age, provided the spouse was married for at least one year. If the deceased died as a result of a work-related accident, occupational disease, or tuber- culosis, spouses are entitled to survivor benefits, regardless of age or the number of years of marriage, provided that the spouse's earnings are sub- ject to mandatory insurance coverage and represent less than 25 percent of the average gross wage. Table 7.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in Romania under the First-Pillar Earnings-Related Scheme Contribution Partial Vesting period rate Eligibility Benefit rate pension Under age 25: 5 years No specific At least 50 Calculated on Depending Age 25­31: 8 years contribution percent loss the basis of on degree Age 31­37: 11 years rate for in capacity number of of disability, Age 37­43: 14 years disability to work points pensioners Age 43­49: 18 years benefits receive 0.75, Age 49­55: 22 years 0.60, or 0.40 Over age 55: 25 years points per year Source: European Commission 2007. 222 Table 7.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions under Romania's First-Pillar Earnings-Related Scheme Orphan Remarriage replacement Total family Eligibility Spouse replacement rate Benefit duration test Orphan age limit rate benefit Eligibility of deceased 50 percent of deceased's For life, if spouse No 16 (26 if orphan 50 percent if 75 percent for for an old-age or a pension if married for meets one of the is student; for sole survivor two survivors; disability pension 15 years; 0.5 percent for first three spouse duration of disability 100 percent for each month less than replacement rate if orphan becomes three or more 15 years up to a minimum conditions; until disabled while survivors of 10 years; 50 percent of youngest child receiving survivor deceased's pension if spouse turns 7; or 6 months benefit) is disabled and married for if none of spouse at least one year; 50 percent replacement rate if spouse has children under conditions is met age 7 Source: European Commission 2007. Romania 223 For spouses who meet none of the three eligibility conditions specified in table 7.6, benefits are paid for six months or until the spouse's youngest child turns seven. Spouses who are eligible for a pension of their own may choose to receive their own pension or a survivor pension. Orphans are entitled to survivor benefits until age 16 (26 if orphan is enrolled in school) or for the duration of their disability in cases in which the orphan becomes disabled while receiving a survivor benefit. The benefit replacement rate for orphans who have lost both parents is 75 percent. Under the second pillar, if a participant dies before becoming eligible for a pension, his or her beneficiaries are entitled to the balance of his or her account. Beneficiaries who are not participating in a private pension fund can elect to receive a lump-sum payment or periodic payments for up to five years. Beneficiaries who are participating in a private pension fund can elect to have the deceased's account merged with their own. Assessment of the Performance of Romania's Pension System The World Bank has established four principles for evaluating public pen- sion systems, which together should guide the process of pension reform (see Holzmann and Hinz 2005). Broadly speaking, these principles include the adequacy and security of benefits, the affordability of contributions, the sustainability of the system over time, and the robustness of the sys- tem in the face of demographic changes and macroeconomic shocks. This chapter focuses primarily on the adequacy of benefits and financial sustain- ability of the first-pillar earnings-related pension scheme. The remaining principles are mentioned only briefly. Adequacy is analyzed through the lens of net replacement rates. Financial sustainability is evaluated using projections of pension expenditure and revenues. Benefit Adequacy Replacement rates are a useful yardstick for measuring the adequacy of pension benefits, because they express benefits relative to preretirement earnings, thereby indicating the degree to which income is replaced when workers retire. Two variants are commonly used. Gross replacement rates compute income replacement as the ratio of benefits paid to pretax pre- retirement earnings. Net replacement rates compute income replacement as the ratio of benefits received (that is, after the payment of taxes and other levies, including contributions for social insurance) to posttax pre- retirement earnings. In general, net replacement rates are a more useful 224 Adequacy of Retirement Income after Pension Reforms measure of benefit adequacy, because they capture the degree to which actual take-home pay is replaced when workers retire. The level of income replacement at retirement is not the only measure of benefit adequacy. For a full assessment of benefit adequacy, it is also important to determine how postretirement indexation rules will affect replacement rates during retirement. Pension benefits in retirement are expected to be indexed to inflation, so that their real value is maintained. In a growing economy with rising real wages, however, mere price index- ation of pensions leads to a deterioration of the relative consumption position of the retirees. For this reason, some countries have introduced mixed indexation of pensions that use varying weights of inflation and wage growth in the indexation formula. For an evaluation of the effect of indexation on replacement rates in Romania, the replacement rates are normalized to 100 and the assump- tions for calculating the replacement rates are maintained (that is, infla- tion is 2.5 percent a year and real wage growth is 2 percent a year). The change in the replacement rate is measured in comparison with full wage indexation or the earnings of an active worker. The results of this analysis indicate that the relative income position of a retiree would deteriorate by 5 percent after 10 years in retirement and by 13 percent after 35 years in retirement. This deterioration is much more modest than that of countries that use price indexation for adjust- ing retirement befits. This is because Romania revalues existing pensions from the first pillar on the basis of changes in the point value, which maintains the value of pensions relative to the average wage. The evalua- tion of income replacement that follows considers replacement rates only at retirement; it does not take into account the impact of indexation poli- cies on replacement rates during retirement. Replacement rates are a function of the formula governing pension benefits; an individual's contribution history; and, in the case of net replacement rates, the rules of income tax, social security contributions, and other relevant levies. The benefit formula establishes the degree to which the system redistributes income across individuals of different lev- els of preretirement earnings. Progressive systems provide higher levels of income replacement to people with lower levels of preretirement income. In general, the degree to which a system is redistributive depends on the existence (and value) of flat transfers and minimum pension guar- antees, the degree to which benefits are earnings related, and the exis- tence of ceilings on earnings subject to contributions. An individual's contribution history can be characterized by his or her age of entry into Romania 225 the labor force, contribution density, and decisions regarding the timing of retirement. To some degree, these three factors are influenced by the incentives embodied in the pension system. The tax and contribution sys- tem affects net replacement rates through the progressiveness of the income tax formula, which taxes (higher) income during a worker's active life more than it taxes (lower) pension benefits in retirement. In addition, social security levies (for pensions; unemployment; health care; and, at times, housing and family benefits) are typically reduced or eliminated altogether in retirement. These benefits are particularly important for low- to middle-income groups. Benchmarks need to be established for the evaluation of the adequacy of the income replacement provided by the earnings-related pension schemes. Unfortunately, there is no consensus on what constitutes ade- quacy. According to one widely respected definition, pensions are ade- quate when they are sufficient to prevent poverty among the elderly and provide the vast majority of the population with a reliable mechanism for smoothing income over their lifetime. Even with a definition, how- ever, establishing benchmarks is problematic, because attitudes vary across countries as a result of social and cultural perceptions. Moreover, benchmarks ignore the other factors that affect the welfare of the elderly--and that also vary across countries--including the existence and generosity of health insurance and long-term care, the cost of housing, the structure of traditional living arrangements, the presence of informal intrafamily or intergenerational sources of financial and nonfinancial support, and the availability and security of other mechanisms for peo- ple to save for their own retirement. One reputable nine-country study (OECD 2001) observes that living standards are roughly comparable for people 10 years older than the normal retirement age and people 15 years younger than the normal retirement age when retirees have disposable income equal to roughly 80 percent of the disposable income of working-age people. In part, this is attributable to the fact that retirees have no work-related expenses (they do not have to commute or buy special clothing or uniforms, for example). This finding, however, does not imply that mandatory first- pillar pension schemes should actually target an 80 percent net replace- ment rate. To the contrary, in middle- and high-income countries, one can reasonably expect individuals to save for their own retirement--and the empirical evidence suggests that, in practice, they do so.8 There is also some evidence to suggest that the ratio between pre- and postretirement income is somewhat independent of the income replacement mandate of 226 Adequacy of Retirement Income after Pension Reforms the public pension system. Put simply, individuals tend to save more in countries with more modest mandates (and vice versa). Because Romania has access to relatively well-developed financial markets, it would seem reasonable to expect middle- and higher- income workers to save enough to finance at least 25 percent, if not closer to 50 percent, of this 80 percent income replacement target. Given this, three benchmarks are provided: a 40 percent net replace- ment rate (which implies that individuals would be expected to save enough to finance half of the total income replacement target); a 60 percent net replacement rate (which implies that individuals would be expected to finance a quarter of the target); and an 80 percent net replacement rate (which implies that individuals, most of whom would be low-income earners, would not be expected to contribute anything toward the target).9 In the following analysis, these three benchmarks are used to evaluate the adequacy of benefits in Romania compared with the average net replacement rate observed in 53 coun- tries around the world, the average net replacement rate observed in selected countries in Europe and Central Asia, and the poverty line in Romania. For an estimation of gross and net replacement rates, two critical dimensions--earnings levels and contribution periods--are considered, with the help of the Analysis of Pension Entitlements across Countries (APEX) model.10 This model generates estimates for replacement rates under steady-state assumptions (that is, as if the rules of the reformed pension scheme had been in place over the entire active life of the indi- vidual). Because life expectancies at retirement are projected to increase over time--which will affect the benefits paid by defined-contribution pension schemes--a reference year must be chosen. The year 2040 is used here, because it provides a sufficiently long contribution period to approximate steady-state conditions. The first critical task is to investigate levels of income replacement across a relevant spectrum of income. Income is represented as a percent- age (50­200 percent) of average earnings. The second task is to investigate the impact on income replacement of differences in the duration, timing, and density of an individual's contribution history (density refers to the percentage of time an individual actually contributes over a given period). To facilitate the presentation of these multidimensional results, we com- pute replacement rates as a function of the age at which an individual exits the labor market. They are presented separately for full-career and partial- career workers. Romania 227 Replacement rates for full-career workers Full-career workers are exam- ined first. For the purpose of this analysis, a full career is defined as contin- uous employment from age 20 to the normal retirement age of 65 for men (effective in 2015). Replacement rates are simulated for an unmarried man working a hypothetical career path under the assumption that real wage growth is 2 percent, inflation is 2.5 percent, the rate of return on invested assets is 3.5 percent, and the worker retires at the statutory retirement age. Gross replacement rates clearly show why the earnings-related pension schemes have been described as providing a strong link between benefits and contributions (figure 7.1). Irrespective of income, gross replacement rates are 72.9 percent, of which 45.0 percentage points are provided by the first pillar and 27.9 percentage are provided by the second pillar. The situation does not change significantly when taxes are taken into consid- eration (figure 7.2). As a result of the impact of taxes and contributions, net replacement rates vary somewhat by income. Pensions for full-career workers in Romania can be considered adequate (figure 7.3). Replacement rates for all levels of preretirement income are substantially higher than the highest benchmark (and high- est for middle-income workers), indicating that the pension system is smoothing consumption effectively from work into retirement. The objective of poverty alleviation is also being met, with levels of income replacement for full-career workers in Romania far higher than regional and world averages.11 Figure 7.1 Gross Replacement Rates in Romania, by Income Level 80 70 60 (percent) 50 rate 40 30 replacement 20 10 gross 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (points system) Source: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. 228 Adequacy of Retirement Income after Pension Reforms Figure 7.2 Sources of Net Replacement Rates in Romania, by Income Level 100 90 80 (percent) 70 60 rate 50 40 30 replacement 20 net 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (points system) taxes Source: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. Figure 7.3 Net Replacement Rates for Male Full-Career Workers in Romania, Europe and Central Asia, and the World 120 100 high benchmark 80 (percent) rate 60 middle benchmark 40 low benchmark replacement net 20 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage Romania average Europe and Central Asia average world average poverty line (percentage of average income) Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. Romania 229 Replacement rates for partial-career workers. Not everyone works from age 20 to the statutory retirement age. Many individuals enter and exit the labor force (often at different ages and for different periods of time) and earn different wages while working (figure 7.4). To examine the ade- quacy of benefits for partial-career workers, we consider three stylized cases. These cases include career type A (someone entering the labor force at age 25 who works continuously for a period of years before leaving the workforce at some point between the ages of 50 and 70 and then claims a benefit); career type B (identical to career type A, except that the worker enters the workforce at age 30 and leaves no earlier than age 55); and career type C (identical to career type A, except that the individual contributes in only three years out of four while in the labor force). In cases where the withdrawal from the formal labor market occurs before the statutory retirement age, the pension is claimed (and the replacement rate calculated) only at the later age. For withdrawals after the statutory retirement age, the ages coincide. Four conclusions can be drawn from examination of net replacement rates for middle-income partial-career workers.12 First, almost all workers Figure 7.4 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Romania, by Career Type and Exit Age 120 100 high benchmark 80 (percent) world and Europe and Central Asia average rate middle benchmark 60 low benchmark 40 poverty line (percentage replacement 20 of average income) net 0 45 50 55 60 65 70 exit age from labor market career type A: career type B: career type C: entry age--25, entry age--30, entry age--25, contribution contribution contribution density--100% density--100% density--75% Source: Authors' calculations based on World Bank 2007a and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. See text for descriptions of career types. 230 Adequacy of Retirement Income after Pension Reforms receive levels of income replacement higher than the poverty line. Second, leaving the workforce very early can be costly. Someone retiring long before reaching the retirement age will receive levels of income replacement barely higher (and, in some cases, lower) than the lowest of the three benchmarks. Third, entering the workforce later in life is costly. Someone entering the workforce at age 30 receives a net replacement rate that is 11­13 percentage points lower than someone entering the workforce at age 25. Fourth, working intermittently is costly. Someone entering the workforce at the same age but who contributes only three years out of four will receive a net replacement rate that is 11­29 percent- age points lower than someone who contributes continuously. Although career type A workers can attain the 80 percent benchmark before reach- ing the normal retirement age, career type B workers must work three years beyond the normal retirement age in order to attain this benchmark. Career type C workers will not be able to attain the 80 percent benchmark even if they work until age 70. Fiscal Sustainability The sustainability of a pay-as-you-go first-pillar pension scheme is best evaluated in actuarial terms by estimating the scheme's actuarial deficit as the difference between its assets and liabilities. If a large actuarial deficit exists, the scheme is financially unsustainable and needs policy actions that increase its assets, reduce its liabilities, or both. A good proxy for the actu- arial deficit is the difference between the present value of the scheme's expected future revenues (that is, contributions and other income) and expected future expenditures (that is, benefit payments, administrative costs, and other expenses) over an extended projection period. The difference between these two values represents an unfunded liability (sometimes referred to as a financing gap) on the public-sector balance sheet. Because this study is also concerned with the time path of revenues and expenditures (and the resulting balance across the projection period ending in 2050), this more pragmatic approach has been taken. Projections of expenditures, revenues, and deficits are presented on the basis of avail- able postreform fiscal projections. Despite (and, in part, because of) the government 's reforms, the first- pillar scheme is projected to continue generating deficits, which are expected to grow for the next three decades relative to GDP before improving slightly (figure 7.5). Rising deficits are caused partly by the need to finance the transition to the second pillar. Revenues are projected to decline steadily, from 6.6 percent of GDP to 3.4 percent of GDP by 2050, as the number of contributors declines and an increasing share of Romania 231 Figure 7.5 Projected Fiscal Balance of Romania's Public Pension System after Reform, 2008­50 15 10 GDP of 5 0 percentage ­5 ­10 2008 2013 2018 2023 2028 2033 2038 2043 2048 expenditures revenues deficit Source: Unpublished World Bank Pension Reform Options Simulation Toolkit (PROST) projections. contributions is diverted from the first to the second pillar.13 Over the same period, expenditures are projected to increase from 7.2 percent of GDP in 2008 to 9.6 percent of GDP by 2050 as the number of benefici- aries increases and benefits are indexed to wages. The net result is a projected deficit of 6.2 percent of GDP in 2050. These projected deficits are also driven by the aging of the popula- tion. Romania's old-age dependency ratio is projected to increase from 23.6 percent in 2008 to 55.3 percent by 2050 (figure 7.6). The aging of the population, in turn, will raise the system dependency ratio (the number of people receiving a pension divided by the number of people contributing to the pension scheme) from 56.9 percent in 2008 to 95.9 percent by 2050.14 What options exist for restoring the system to fiscal balance? Unfor- tunately, for policy makers, the options are limited. Revenues can be increased by increasing the contribution rate. Alternatively--or in addi- tion, because the options are not exclusive--expenditures can be reduced by cutting benefits, increasing the minimum number of years required to become eligible for benefits, or delaying the payment of benefits by raising the retirement age further. Because raising the con- tribution rate could threaten competitiveness and will likely strengthen incentives for tax evasion, it is typically not embraced. Raising the con- tribution rate would also represent a reversal of policy because Romania has been deliberately reducing the rate to dampen the adverse impact of high taxes on labor markets. This leaves policy makers with limited 232 Adequacy of Retirement Income after Pension Reforms Figure 7.6 Projected Old-Age and System Dependency Ratios in Romania, 2008­50 120 100 80 60 (percent) ratio 40 20 0 2008 2010 2020 2030 2040 2050 year system dependency ratio old-age dependency ratio Sources: Unpublished World Bank Pension Reform Options Simulation Toolkit (PROST) projections; Reiterer 2008. options: cutting benefits, tightening eligibility conditions, or raising the retirement age. Given that a major part of the deficit reflects the tran- sition costs associated with the second pillar, the government may also consider financing part or all of these costs using general revenues. If it does otherwise, restoring sustainability may reduce the adequacy of benefits provided to future beneficiaries. If retirement ages are left unchanged and the current structure of the system is retained, further cuts in benefits--on the order of a 59 percent reduction in the average benefit provided under the first pillar--will be required for the system to become sustainable (figure 7.7). If benefits are adjusted to maintain a similar fiscal balance in proportion to the overall size of the first-pillar scheme, full-career workers will receive replacement rates that are roughly 28 percentage points lower in 2050 than they are today. Two conclusions can be drawn from comparing these new (lower) net replacement rates with the three benchmarks. First, a 59 percent reduc- tion in first-pillar benefits would not cause income replacement for full- career workers to fall below the poverty line, except for those with very low incomes. This indicates that Romania's pension system would still broadly achieve its poverty alleviation objective. Second, the same reduction in benefits would still support the objective of smoothing Romania 233 Figure 7.7 Net Replacement Rates for Male Workers in Romania before and after Benefit Adjustment 120 100 (percent) 80 high benchmark rate 60 middle benchmark 40 low benchmark replacement net 20 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage net replacement rate adjusted net replacement rate poverty line (percentage of average income) Source: Authors' calculations based on World Bank 2007a and the APEX model. consumption for middle- and high-income full-career workers, because levels of income replacement are still equal to or higher than the mid- dle 60 percent benchmark. Replacement rates for low-income workers, however, would fall substantially below the 80 percent benchmark. This last observation is subject to three caveats. First, this analysis con- siders only full-career workers, while the average worker now contributes for only about 27­30 years, substantially less than the 45 years of a full career. Contributing to the pension scheme for only 35 years, for exam- ple, reduces net income replacement 23 percentage points for middle- income workers. This suggests that restoring the system to fiscal balance on the basis of benefit cuts alone may not provide many partial-career workers with adequate levels of income replacement. Second, if benefit cuts are combined with further increases in the retirement age, the ben- efit cuts will not need to be as steep to restore fiscal balance. Third, work- ers always have the option of saving outside of the first-pillar pension scheme. To increase income replacement by 1 percent, for example, a full- career worker would need to save only about 0.43 percent of his or her earnings from age 40 to the current age of retirement.15 234 Adequacy of Retirement Income after Pension Reforms Conclusions Recognizing that a pension system that offers generous benefits, low retirement ages, and lax benefit eligibility conditions would not be sus- tainable over the long term as the population ages, Romania's govern- ment introduced substantial pension system reforms in 2000. These reforms raised retirement ages, extended the service period required to become eligible for a full pension, imposed new conditions on early retirement, and replaced the traditional defined-benefit formula with a new formula based on points. In 2006, the government also passed legis- lation to introduce a mandatory, fully funded, defined-contribution scheme, which became operational in May 2008. The resulting gross and net replacement rates for full-career workers after reform are projected to be well above regional and world averages, especially for middle- and high-income workers (see chapter 1). Future net replacement rates for full-career workers are projected to be 88­97 percent across the analyzed income spectrum. The generosity of these initial replacement rates is enhanced by generous indexation poli- cies, which will fully index first-pillar benefits to wages. As in other coun- tries, workers with less than full careers--because they left the workforce before reaching retirement age, worked intermittently, or have gaps in their employment history--risk receiving a level of income replacement closer to the lower benchmark of 40 percent or possibly even lower. Following reform, Romania's pension system is projected to generate ever-larger deficits relative to GDP for the next three decades before improving slightly to 6.2 percent of GDP by 2050. These deficits are driven by the aging of the population, as well as by the generous benefits in the reformed system (especially given recent large increases in benefits) and the transition costs associated with replacing part of the unfunded first-pillar scheme with a funded second-pillar scheme. Population aging by itself will reduce the number of contributors relative to the number of beneficiaries to such an extent that by 2050 the number of pension system beneficiaries will approach the number of contributors. The lower retirement age applied to women, in combination with their longer life expectancy, greatly increases their lifetime benefit costs rela- tive to their contributions. To restore long-term fiscal balance to the first-pillar scheme without recourse to general revenue financing, the government needs to raise retirement ages further--rough estimates suggest to well above age 70 by 2050. To realize the full fiscal impact of this measure, the government Romania 235 must maintain income replacement at levels associated with current retirement ages. Increasing the retirement age in step with increases in life expectancy at retirement is a natural choice, for both individuals and for policy makers, but it requires cross-sectoral policy reforms to enable eld- erly workers to continue to participate in the labor market.16 Other options for restoring fiscal balance including cutting first-pillar benefits at retirement, reducing the generosity of benefit indexation, or adopting some combination of the two. Rough estimates suggest that aver- age first-pillar replacement rates would have to fall by about 28 percentage points. Moving from wage to price indexation would reduce average first- pillar benefits by some 25 percentage points. Reducing benefits by this amount would not compromise the objective of alleviating poverty among the elderly, because levels of income replacement would still not fall any- where near the poverty line. Doing so may, however, compromise the objective of smoothing lifetime consumption for full-career workers of all income levels when measured against the highest of the three benchmarks examined here. Individuals who wish to defer more of their lifetime con- sumption into retirement would still have the option, of course, of partici- pating in the voluntary third-pillar pension scheme. Notes 1. The required contribution rate under the old system would have needed to be increased from 25.5 percent to 50 percent in the long term for the sys- tem to provide the same level of benefits provided now (World Bank 1998). 2. For more information regarding these population projections, see Reiterer (2008). 3. In 2007, the average number of points of existing pensioners was 1.38. 4. As a percentage of the gross average wage, the point value was about 31.0 percent in 2006, 37.5 percent in 2007, and 37.5 percent in 2008; it will rise to 45 percent in 2009. Existing pensions are revalued on the basis of changes in the point value by multiplying an individual's average points by the new point value. The increase in the point value over the past few years resulted in pensions rising faster than wages, because the point value, as a share of the average wage, increased from 30 percent to 45 percent (in 2009)-- an increase of 50 percent. 5. In 2007, life expectancy at retirement age was 13.6 years for men and 20.9 years for women. These values are projected to increase to 16.0 years for men and 24.4 years for women by 2050. Lengthening life expectancy will substantially increase the average period over which benefits are paid. Many 236 Adequacy of Retirement Income after Pension Reforms pension systems worldwide are designed to provide individuals with 15 years of benefits in retirement (Schwarz 2006). 6. The law does not specify by how much first-pillar benefits will be reduced when contributions are made to the funded second-pillar scheme. This study assumes that the reduction will be proportional to the reduced share of con- tributions flowing to the first pillar. 7. The second-pillar pension scheme will eventually provide these benefits as well; many of the provisions governing the payment of benefits have yet to be determined. 8. In Chile, for instance, 70 percent of retirees from the mandatory public pen- sion system own their home, which is a form of savings (Valdés-Prieto 2008). 9. These benchmarks approximate the standards developed by the International Labour Organization (ILO) (1952) and the Council of Europe (1990). ILO Convention 102 of 1952 sets a minimum benefit equal to 40 percent of the reference wage for married men of pensionable age. This amount was raised to 45 percent in 1968. The European Code of Security of 1990 sets a mini- mum standard for members of the Council of Europe equal to 65 percent for married people of a specific age. 10. The APEX model was developed by Axia Economics, with funding from the Organisation for Economic Co-operation and Development and the World Bank. The model codes detailed eligibility and benefit rules for first- and sec- ond-pillar schemes based on available public information that has been ver- ified by country contacts. Because the details of the rules sometimes change on short notice (and limited public disclosure), the calculations presented here should be considered as best approximations only. 11. As a proxy for the poverty line, a figure of 35 percent of the average net wage is used, because this percentage broadly approximates a US$2.25-a-day poverty line converted into national currency, adjusted for purchasing power parity, expressed relative to the national average net wage, and averaged across the eight study countries. 12. Only middle-income, partial-career workers are examined because replace- ment rates are comparable for workers with lower and higher levels of pre- retirement income. 13. Projections reflect the 1.5 percentage point decrease in the contribution rate planned by the government as well as the diverting of contribution revenues to the second pillar. 14. Over the same period, the population is projected to decrease from 21.5 mil- lion to 17.1 million (Reiterer 2008). 15. This estimate is based on the assumption that real wage growth is 2.0 per- cent, the net real rate of return on invested assets is 3.5 percent, and bene- fits (from both the unfunded and the funded pillars) are price indexed. Romania 237 16. See Holzmann, MacKellar, and Repansek (2009) for a conference volume that addresses these issues for the countries of southeastern Europe. Bibliography Council of Europe. 1990. European Code of Social Security (Revised). Rome. European Commission. 2007. Mutual Information System and Social Protection (MISSOC) database. http://ec.europa.eu/employment_social/. GVG (Gesellschaft für Versicherungswissenschaft und -gestaltung) and European Commission. 2003. Study on the Social Protection Systems in the 13 Applicant Countries: Romania Country Report. Cologne. Holzmann, R., and R. Hinz. 2005. Old-Age Income Support in the 21st Century. Washington, DC: World Bank. Holzmann, R., L. MacKellar, and J. Repansek, eds. 2009. Pension Reform in South- eastern Europe: Linking to Labor and Financial Market Reforms. Washington, DC: World Bank. ILO (International Labour Organisation). 1952. ILO Convention 102. Geneva: ILO. ------. 1967. ILO Convention 128. Geneva: ILO. OECD (Organisation for Economic Co-operation and Development). n.d. Database. http://www.oecd.org/dataoecd/13/30/38708660.pdf. ------. 2001. Ageing and Income: Financial Resources and Retirement in 9 OECD Countries. OECD: Paris. Reiterer, A. 2008. Population Development and Age Structure in Southeastern Europe until 2050. World Bank, Washington, DC. Schwarz, A. 2006. Background paper for Public Expenditure Review: Overview of Pension Sector in Serbia. World Bank, Washington, DC. U.S. Social Security Administration. 2006 Social Security Systems throughout the World: Europe. Washington, DC: Social Security Administration. Valdés-Prieto, S. 2008. Designs for the First Pillar Pensions and the 2008 Chilean Reform. http://editorialexpress.com/cgi-bin/conference/download.cgi?db_ name=SECHI2008&paper_id=130. Whitehouse, E. 1999. Tax Treatment of Funded Pensions. Washington, DC: World Bank. World Bank. 1998. Romania: Pension Reform Note. Washington, DC: World Bank. ------. 2003. Romania Poverty Assessment. Washington, DC: World Bank. ------. 2004. The Pension System in Romania: Challenges of Pursuing an Integrated Reform Strategy. Washington, DC: World Bank. ------. 2006. Background paper for the Romania Poverty Assessment. World Bank, Washington, DC. 238 Adequacy of Retirement Income after Pension Reforms ------. 2007a. Pensions Panorama. Washington, DC: World Bank. ------. 2007b. Background paper for the Romania Poverty Monitoring Analytical and Advisory Assistance Program. World Bank, Washington, DC. WHO (World Health Organization). 2000. Healthcare Systems in Transition: Romania. Copenhagen: WHO. ------. 2008. Database. http://www.who.int/research/en/. C H A P T E R 8 The Slovak Republic Following the dissolution of Czechoslovakia in 1993, the Slovak Republic inherited a public pension system financed on a pay-as-you-go basis (meaning that contributions from current workers are used to pay benefits to current beneficiaries). The pension system effected considerable redis- tribution of income from the comparatively well-off population to those less fortunate. This loose connection between contributions and benefits, combined with the increased informality in labor markets, caused revenues to gradually decline, from an amount equivalent to 8 percent of gross domestic product (GDP) in the mid-1990s to about 7 percent by 2002. With expenditures hovering around 7.3 percent of GDP during this period, the fiscal condition of the pension system began to deteriorate. In 1999, the system experienced its first deficit. Projections suggested the pension scheme would face even greater fiscal challenges in the future, as a result of the rapidly aging population, with deficits eventually reaching an amount equivalent to 10 percent of GDP. Recognizing the need for reform, the government introduced major systemic changes to the existing pay-as-you-go scheme in 2004 intended to arrest growing deficits and restore fiscal balance. In 2005, it introduced a privately managed, fully funded defined-contribution scheme. Together, these reforms have considerably improved the fiscal health of the pension 239 240 Adequacy of Retirement Income after Pension Reforms system. Projections suggest that deficits will reach 4.4 percent of GDP by 2050--less than half their previous level. The aging of the Slovak population will almost certainly compel the government to enact further reforms. Further improvements will require that benefits be made less generous or retirement ages be raised further (or some combination of both), a tradeoff that will become even more pronounced as people live longer. Against this backdrop, this chapter evaluates the Slovak Republic's pension system, focusing on fiscal sustainability and benefit adequacy. Adequacy is evaluated through the lens of statutory net replacement rates for different retirement ages, patterns of contributions, and income levels relative to international benchmarks. This chapter is organized as follows. The next section discusses the motivation for the reforms. The following section describes the key char- acteristics of the reformed pension system. The third section assesses the adequacy of pension benefits and the fiscal sustainability of the system. The last section draws conclusions. Motivation for Reform Following independence from Czechoslovakia, the Slovak Republic under- went a difficult transition from central planning to a market economy. The difficulty of this transition was reflected in the fiscal balance of the pen- sion system, which experienced declining contribution revenues (as a result of increasing informality in the labor markets, higher formal-sector unemployment, and stagnant--at times falling--average wages) and rela- tively constant expenditure (because the number of beneficiaries remained relatively flat).1 By 1999, the pension system began generating deficits (table 8.1). Although these deficits were not huge relative to GDP, projections sug- gested they would grow to unaffordable levels as a result of the aging of the population. By 2050, revenues were expected to hover around 6.5­7.0 percent of GDP while expenditures were expected to increase to 16.4 percent of GDP, resulting in deficits of roughly 10 percent of GDP (figure 8.1). For Slovak policy makers, this was the impetus for reform. The aging of the Slovak population was largely behind these projec- tions. The old-age dependency ratio (the population age 65 and older divided by the population age 20­64) was projected to increase from 18.3 percent in 2005 to 54.9 percent by 2050 (Reiterer 2008) (figure 8.2). The system dependency ratio was projected to increase from 52 percent The Slovak Republic 241 Table 8.1 Fiscal Balance of the Slovak Republic's Pension System before Reform, 1995­2002 (percentage of GDP) Year Revenues Expenditures Balance 1995 7.8 7.3 0.5 1996 8.1 7.3 0.8 1997 7.3 7.2 0.1 1998 7.3 7.3 0.0 1999 6.8 7.4 ­0.6 2000 7.3 7.5 ­0.2 2001 6.9 7.4 ­0.5 2002 6.9 7.3 ­0.4 Source: World Bank 2004. Figure 8.1 Projected Fiscal Balance of the Slovak Republic's Public Pension Scheme before Reform, 2000­50 20 15 10 GDP of 5 0 percentage ­5 ­10 ­15 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 expenditures revenues deficit Source: Unpublished World Bank Pension Reform Options Simulation Toolkit (PROST) simulations. in 2005 to 102 percent by 2050, suggesting that the number of benefici- aries would eventually exceed the number of contributors. Projections suggested that 2.4 contributors were needed per beneficiary for the scheme to be sustainable (World Bank 2004). To address the problem, in 2004 the government redesigned the parameters of the existing pay-as-you-go system by switching from a defined-benefit formula for computing pension benefits to a system based on points. Under a points system, individuals are awarded 242 Adequacy of Retirement Income after Pension Reforms Figure 8.2 Projected Old-Age and System Dependency Ratios in the Slovak Republic before Reform, 2005­50 120 100 80 60 percent 40 20 0 2005 2010 2020 2030 2040 2050 system dependency ratio old-age dependency ratio Sources: Reiterer 2008; unpublished World Bank Pension Reform Options Simulation Toolkit (PROST) simulations. points for each year they contribute, where points are a function of the ratio of their earnings relative to the economywide average wage. At retirement, benefits are based on the total number of points accu- mulated. In 2005, the government introduced a mandatory fully funded defined-contribution scheme, whereby benefits depend on the amount of an individual's contributions and the rates of return earned on invested assets. Characteristics of the Slovak Republic's Pension System This section describes the main characteristics of the Slovak Republic's pension system. They include the design of the individual pillars of social insurance; the rules governing pension system taxation, institutional structure, and coverage; and the provisions governing old-age, disability, and survivorship pensions. The design of the pension system is assessed using a conceptual framework developed by the World Bank, which gen- erally recommends including a funded component if conditions are appropriate but increasingly recognizes that a range of choices is available to policy makers to provide effective old-age protection in a manner that is fiscally responsible (see Holzmann and Hinz 2005). The Slovak Republic 243 In general, the World Bank supports pension systems composed of some combination of five basic pillars: · a noncontributory (or zero) pillar (in the form of a demogrant, social pension, or social assistance benefit) intended to provide a minimal level of income protection; · a first-pillar contributory system linked to earnings, which seeks to replace a portion of preretirement income; · a mandatory second pillar (essentially, individual savings accounts), which can be designed in various ways; · a voluntary third pillar, which is flexible and discretionary (this pillar, too, can take a variety of forms); and · a fourth pillar of informal intrafamily or intergenerational sources of financial and nonfinancial support to the elderly, including access to health care and housing. Pillar Design The design of the Slovak pension system incorporates all five of the pillars recommended by the World Bank (table 8.2). The publicly managed non- contributory zero pillar, financed with general tax revenues, redistributes income to lower-income groups using means testing. The reference point for computing the amount of an individual's benefit is the minimum subsistence level, which implies that the noncontributory benefit is adjusted as the subsistence level is changed. Both the traditional publicly managed pay-as-you-go first pillar and the privately managed, fully funded second pillar are earnings-related schemes. First-pillar benefits are com- puted on the basis of the number of points accumulated over the course of an individual's career. Postretirement benefits are indexed using a combination of wage and price growth (Swiss indexation), whereby benefits increase with wages but at a lower rate. Second-pillar benefits are a function of an individual's contributions and investment earnings. At retirement, account balances are converted into annuities based on the accumulated capital in an individual's account and the individual's condi- tional life expectancy. Supplementing these earnings-related schemes is a voluntary privately managed third pillar, intended to provide individuals with a mechanism for adding to the benefits provided by the mandatory pillars. The fourth pillar provides health care to the elderly as part of the overall health care system. The taxation of contributions and benefits varies across pillars. The zero, first, and fourth pillars are completely exempted from taxation. The 244 Table 8.2 Structure of the Slovak Pension System Taxation Investment income/ Benefit capital Scheme type Coverage Type Function Financing Generic benefit indexation Contributions gains Benefits Zero pillar Universal Means Redistributive Tax revenues Difference Minimum n.a. n.a. Exempt (public tested between minimum subsistence noncontributory) subsistence level (close to level and actual consumer price income index) First pillar (public, Mandatory Points Insurance Percentage of Based on the number 50 percent Exempt n.a. Exempt earnings related) individual of points earned inflation, 50 earnings percent nominal wage growth Second pillar Mandatory Defined Insurance Percentage of Pension from capital Depends on Exempt Taxed Exempt (private, earnings contribution individual accumulation options chosen related) earnings Third pillar (private, Voluntary Defined Insurance Voluntary Pension from capital Depends on Exempta Exempt Taxed voluntary) contribution contributions accumulation options chosen Fourth pillar Mandatory n.a. Insurance Percentage of Basic health service n.a. Exempt n.a. n.a. (public individual package health care) earnings plus tax revenues Sources: European Commission 2007a, 2007b; World Bank 2004. n.a. = Not applicable. a. An amount up to 12,000 koruny annually is exempt from taxation. The Slovak Republic 245 second pillar is subject to an exempt-taxed-exempt regime, meaning that contributions are exempt from taxation, investment income is taxed, and benefits are exempt. The third pillar is subject to a classic expenditure tax (exempt-exempt-taxed) in which contributions (up to 12,000 koruny [SK] annually) and investment returns are exempt from taxation while benefits are taxed. (For a discussion of the taxation of retirement savings, see box 1.1 in chapter 1.) Noncontributory scheme. In the Slovak Republic, the elderly are eligible for a noncontributory benefit as part of a general program of social assis- tance designed to provide a minimum level of income protection to the overall population irrespective of age. Both elderly pensioners and elderly people who are ineligible for a pension may apply for assistance. The amount of the benefit varies as a function of household income and size: higher benefits are paid to households with no other sources of income, while lower benefits are paid to households with modest income from other sources. The computation of the social assistance benefit for eld- erly people who contributed to the pension system for at least 25 years takes into account only 75 percent of their pension, with an additional 1 percent decrease with each additional year of contributions before retirement. Once a pensioner's benefit has been adjusted, the resulting figure is increased to ensure that total household income meets the mini- mum income thresholds established by law. At the end of 2007, social assistance benefits were paid to 182,479 beneficiaries, at a cost equivalent to 0.45 percent of GDP. There were 38,606 beneficiaries of pensionable age (21.2 percent of all beneficiaries). Earnings-related schemes. Before reform, the Slovak pension system was highly redistributive, partly as a result of a cap on the accumulation of ben- efits. Although redistribution in a pension system is not necessarily a design flaw, the cap created an incentive for workers and employers to underre- port earnings (which may have contributed to the fiscal problems of the pension system in the late 1990s, during the country's transition to a mar- ket economy). The reforms of 2004 eliminated the minimum pension, con- sistent with the separation of social insurance from social assistance. Social assistance benefits are now used to provide any needed redistribution, while the social insurance program is used as a contribution-based instrument of savings (World Bank 2004). Under the reformed system, first-pillar benefits are now computed on the basis of the number of points accumulated throughout an individual's 246 Adequacy of Retirement Income after Pension Reforms career, a change that ensures that benefits reflect lifetime wages. A point system effectively mimics a notional account approach to pension reform.2 The former system, which used a complicated (and highly redis- tributive) formula based on the individual's best years of wages, was abol- ished. Contributions are now paid up to a wage of four times the average wage; benefits are computed on the basis of wages up to three times the average wage (under the old system benefits were based on a number of best years' wages).3 Benefits are actuarially reduced for early retirement and increased for delayed retirement. Postretirement benefits are indexed using Swiss indexation. These reforms improved transparency and tight- ened the link between contributions and benefits, which reduced incen- tives for noncompliance. Before reform, retirement ages were very low: life expectancy at retire- ment age was 16.8 years for men and 25.1 years for women (unpublished World Bank Pension Reform Options Simulation Toolkit [PROST] simu- lations). To combat the pressures created by these low retirement ages and gradually increasing life expectancies, retirement ages are gradually being increased to age 62 for both men and women (table 8.3). This will reduce the average period of benefit collection to 15.7 years for men and 20.8 years for women. The change will reduce benefit costs and increase pen- sion system revenue, because workers will contribute to the scheme for more years. Even with these changes, however, retirement ages remain low, especially for women given their substantially higher life expectancy. In 2005, the government introduced a mandatory fully funded defined-contribution scheme, consistent with its broader objective of moving away from a pension system based entirely on a single pillar.4 In the new two-pillar structure, half of the contributions go to the newly established second scheme while the other half continue to finance first-pillar benefits. Voluntary scheme. In 2005, the voluntary third-pillar pension scheme (pre- viously available only to workers with an established employee­employer relationship) was opened to people over age 18 (table 8.4). Benefits are awarded to people over age 55 after a minimum of 10 years of contribu- tions. Participation is encouraged through tax policy. Contributions made by employees to the scheme are deductible from personal income taxes up to an annual ceiling of Sk 12,000. Contributions made by employers on behalf of their employees are deductible from enterprise taxes up to 6 per- cent of wages. Part of an individual's account balance can be distributed as a lump sum upon reaching retirement, provided certain conditions are met. Table 8.3 Parameters of Earnings-Related Schemes in the Slovak Republic before and after Reform Pension Vesting Contribution assessment Scheme type Period period Contribution rate ceiling Benefit rate base Retirement age First pillar Prereform 25 years 28 percent (21.6 Sk 32,000 2 percent Highest 5 60 for men, 53­57 (earnings related) percent by employer, accrual rate of last 10 for women 6.4 percent by for first 25 years depending, on employee) years, 1 average number of percent for children next 17 years Postreform 15 years 18 percenta 3 times average 1.19 percent Lifetime 62 for all (men by (4 percent by wage per year average 2007, women by employee, 14 indexed to 2016) percent by employer) nominal wage growth Second pillar Prereform n.a. n.a. n.a. n.a. n.a. n.a. (earnings related) Postreform 10 years 9 percent by employer 3 times average Pension from Accumulated 62 for all wage capital funds accumulation Sources: European Commission 2007a, 2007b; World Bank 2004. n.a = Not applicable. a. The total contribution for the old-age, disability, and survivors pensions reserve fund is 28.75 percent (18 percent for old-age reserve fund, 6 percent for disability reserve fund, and 4.75 percent for employer reserve fund). Individuals participating only in the first pillar pay 18 percent to the first pillar for old-age coverage (14 percent by employer, 4 by employee). Individuals participating in both the first and the second pillars pay 9 percent (5 percent by employer, 4 percent by employee) to the first pillar and 9 percent (entirely employer) to the 247 second pillar for old-age coverage. 248 Adequacy of Retirement Income after Pension Reforms Table 8.4 Characteristics of the Voluntary Scheme in the Slovak Republic Lump-sum Tax advan- Contributions payments Vesting Retirement tages to par- tax deductible possible in Coverage period age ticipants by employers retirement Anyone over age 18 10 years 55 Yes Yes Yes Source: European Commission 2007b; U.S. Slovak Embassy Web site (http://www.slovakembassy-us.org). In the event of disability, the account balance, including any investment earnings, is distributed to the participant or, in case of death, to his or her survivors. In 2006, assets of the scheme amounted to 1 percent of GDP. Some 673,352 people (17.4 percent of the working-age population) partic- ipated (Cillíková 2006). Health care system. Health care in the Slovak Republic is provided pri- marily through the mandatory health insurance scheme, administered by one of five health insurance companies. It is financed primarily by contributions and by copayments for some medical services and phar- maceuticals. Voluntary health insurance is also available, although par- ticipation is negligible. Contributions to the mandatory health insurance scheme from the economically active population constitute about 70 percent of the total; contributions from the government on behalf of the nonactive popula- tion (including pensioners) constitute 30 percent. The contribution rate of 14 percent (10 percentage points of which are paid by employers and 4 percentage points by employees) is levied on income between the minimum wage and a ceiling of three times the average wage. The gov- ernment contributes 4 percent of the average wage on behalf of the nonactive population, which represents 60 percent of the total popula- tion (World Bank 2002). All insured people, including the elderly, have access to the same basic health care benefits specified by law. In 2005, health expenditure accounted for 7.0 percent of GDP, 74.4 percent of which was public expenditure and 25.6 percent was private expenditure. Of the private expenditure, 88.1 percent was attributable to out-of-pocket expenditure (informal payments, direct payments, and copayments) (WHO 2008). Institutional Structure and Coverage of Earnings-Related Schemes The mandatory pillars of the Slovak pension system cover all salaried employees and the self-employed, including farmers. Voluntary partici- pation is open to individuals older than age 16 and to the self-employed The Slovak Republic 249 who earn less than the minimum wage. Special provisions are in place for the police corps, the Slovak information service, the national secu- rity authority, the prison and justice guard corps, railway police, and customs officers. The Social Insurance Agency is responsible for administering the public system. It collects contributions for both the first and the second pillars, act- ing as a clearinghouse (meaning that it transfers second-pillar contributions to designated pension-fund management companies, thereby reducing the logistical burden on employers). Roughly 85.5 percent of the labor force (59.4 percent of the working-age population) contributes to the first pillar, of which 70 percent also participate in the second pillar. In 2006, six pension-fund management companies were operating in the mandatory (second-pillar) pension fund market and four companies were operating in the voluntary (third-pillar) market. Private pension- fund management companies must be licensed by the National Bank of Slovakia, which oversees them. Pension fund management companies are required to offer three funds--a conservative fund, a balanced fund, and a growth fund--each governed by different portfolio investment guidelines. The conservative fund is limited to investing in bonds and money market instruments. The balanced fund can invest up to 50 percent of its assets in equities, while the growth fund can invest up to 80 percent in equities. Rules on external investment are liberal: only 30 percent of investments must remain in the country. Even this requirement is under review as potentially in conflict with European Union rules on the free flow of cap- ital among member countries. Structure of Benefits The earnings-related pension schemes provide old-age, disability, and sur- vivorship pensions. The provisions governing each of these types of ben- efits are discussed as follows. Old-age benefits. In the first-pillar scheme, participants must have at least 15 years of service and must reach the minimum retirement age to be eligible for an old-age pension. Retirement ages for men were increased from 60 to 62 between 2004 and 2006, at the rate of 9 months per year. Under the old rules, the retirement age for women ranged from 53 to 58 (as a function of the number of children). Under the new rules, the age is being raised at the rate of 9 months per year, such that the retirement age for all women, regardless of the number of children, will be 62 by 2015. 250 Adequacy of Retirement Income after Pension Reforms Benefits from the first pillar are computed on the basis of a point system. People enrolled in the reformed first pillar (but not the second pil- lar), most of whom are older, accrue benefits at the rate of 1.19 percent per year of service. People enrolled in both the reformed first pillar and the new second pillar accrue benefits under the first pillar at half the rate (meaning their accrual rate is 0.6 percent); the remainder of their benefits come from their individual investment account (only half of their contri- butions are used to pay for the cost of their benefits under the first pillar). People with at least 15 years of contributions may retire at any age before reaching the minimum retirement age, but their benefits are penalized by 6 percent per year until they reach the minimum retirement age. Early retirement is not permitted if the resulting pension is less than 1.2 times the subsistence income. Retirement can be deferred after a person reaches the minimum retire- ment age. Benefits are increased by 6 percent for each year retirement is deferred. For pensioners who continue to work or who reenter the work- force after retiring once, the pension is recalculated when they retire the second time, with half of the points earned since their first retirement credited toward their pension. Participants in the second pillar may elect to have some of the pro- ceeds of their account paid in a lump sum or as a programmed with- drawal over a prespecified time period, as long as the remainder of their funds are sufficient to purchase a life annuity from an insurance company equal to or larger than 60 percent of the subsistence minimum. In the case of programmed withdrawals, the pension fund company must con- tinue to invest assets in the account. Each year, as a function of the returns earned on invested assets, a monthly withdrawal amount is computed such that the account will be completely exhausted by the end of the term of the contract. Disability benefits. Disability benefits are available to both individuals enrolled only in the first-pillar scheme and to those enrolled in both the first- and second-pillar schemes (table 8.5). Disabled people enrolled only in the first pillar are given credit for years lost to disability. Benefits are based on the worker's average wage at the time of disability. For people with 40­70 percent impairment, the pension is reduced by the extent of their disability. When the individual reaches the minimum retirement age, the disability pension is replaced by the old-age pen- sion to which he or she would have been entitled. For fully disabled people who do not work after becoming disabled, the two pensions are The Slovak Republic 251 equal, because their disability pension gives them credit for having worked a full career. For disabled people enrolled in both the first and second pillars, since January 2008 benefits have been determined by the first- pillar rules only, with years of disability excluded from the old-age pen- sion calculation under the first pillar. As of January 2008, the disability fund stopped paying contributions to the second pillar, which now provides no disability benefits. Survivor benefits. Survivor benefits are awarded under both the first and second pillars to the dependents of individuals who at the time of their death were receiving (or had met the criteria to receive) an old-age or dis- ability pension (table 8.6). Eligible survivors include widows or widowers and orphaned children. Spouses receive 60 percent of the deceased's Table 8.5 Eligibility Conditions for and Benefits Provided by Disability Pensions under the First-Pillar Earnings-Related Scheme in the Slovak Republic Vesting period Contributions Eligibility Benefit rate Partial pension Under age 20: less 3 percent by At least 40 1.19 percent Pension is than 1 year employer, percent loss per year prorated if Age 20­22: 1 year 3 percent by of capacity to disability is Age 22­24: 2 years employee; work 40­70 Age 24­26: 3 years ceiling of percent Age 26­28: 4 years three times Over age 28: 5 years average wage Sources: European Commission 2007a; World Bank 2004. Table 8.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions under the First-Pillar Earnings-Related Scheme in the Slovak Republic Spouse Orphan Total replacement Benefit Remarriage Orphan replace- family Eligibility rate duration test age limit ment rate benefit Eligibility of 60 percent of For life, if Pension 26 40 percent 100 percent, deceased deceased's spouse is ceases if regardless for old-age pension 70 percent survivor of number or disability disabled, remarries of survivors pension caring for a child, or at retire- ment age; otherwise, for one year Source: European Commission 2007a. 252 Adequacy of Retirement Income after Pension Reforms pension, subject to rules governing remarriage and the duration of ben- efits. Irrespective of the number of orphans, total benefits paid to all survivors cannot exceed the total benefit to which the deceased was orig- inally entitled. Survivors of individuals enrolled in the second pillar are entitled to the entire accumulated balance of the deceased's account. Survivors of old-age pensioners receive 60 percent of the annuity payable to the deceased plus any remaining balance in the deceased's account. Assessment of the Performance of the Slovak Pension System The World Bank has established four principles for evaluating public pen- sion systems, which together should guide the process of pension reform (see Holzmann and Hinz 2005). Broadly speaking, these principles include the adequacy and security of benefits, the affordability of contri- butions, the sustainability of the system over time, and the robustness of the system in the face of demographic changes and macroeconomic shocks. This chapter focuses primarily on the adequacy of benefits and the financial sustainability of the earnings-related pension schemes. The remaining principles are mentioned only briefly. Adequacy is analyzed through the lens of net replacement rates. Financial sustainability is evalu- ated using projections of pension expenditure and revenues. Benefit Adequacy Replacement rates are a useful yardstick for measuring the adequacy of pension benefits, because they express benefits relative to preretirement earnings, thereby indicating the degree to which income is replaced when workers retire. Two variants are commonly used. Gross replacement rates compute income replacement as the ratio of benefits paid to pretax pre- retirement earnings. Net replacement rates compute income replacement as the ratio of benefits received (that is, after the payment of taxes and other levies, including contributions for social insurance) to posttax pre- retirement earnings. In general, net replacement rates are a more useful measure of benefit adequacy, because they capture the degree to which actual take-home pay is replaced when workers retire. The level of income replacement at retirement is not the only measure of benefit adequacy. For a full assessment of benefit adequacy, it is also important to assess how postretirement indexation rules will affect replace- ment rates during retirement. Pension benefits in retirement are expected to be indexed to inflation so that their real value is maintained. In a grow- ing economy with increasing real wages, mere price indexation of pensions, The Slovak Republic 253 however, leads to a deterioration of the relative consumption position of the retirees. Individuals with otherwise identical work histories will receive different pensions depending on when they retire. For this reason, some countries, such as the Slovak Republic, have introduced mixed indexation of pensions with varying weights of inflation and wage growth in the index- ation formula. In order to evaluate the effect of indexation on replace- ment rates in the Slovak Republic, the replacement rates are normalized to 100 and the assumptions for calculating the replacement rates are maintained (that is, inflation is 2.5 percent per year and real wage growth is 2 percent per year). The change in the replacement rate is measured in comparison with full wage indexation or compared to an active worker. The results of this analysis indicate that the relative income position of a retiree would deteriorate by 13 percent after 10 years in retirement and by 37 percent after 35 years in retirement. The evaluation of income replacement that follows considers replacement rates only at retirement; it does not take into account the impact of indexation policies on replace- ment rates during retirement. Replacement rates are a function of the formula governing pension benefits; an individual's contribution history; and, in the case of net replace- ment rates, the rules of income tax, social security contributions, and other relevant levies. The benefit formula establishes the degree to which the system redistributes income across individuals of different levels of preretirement earnings. Progressive systems provide higher levels of income replacement to people with lower levels of preretirement income. In gen- eral, the degree to which a system is redistributive depends on the existence (and value) of flat transfers and minimum pension guarantees, the degree to which benefits are earnings related, and the existence of ceilings on earnings subject to contributions. An individual's contribution history can be characterized by his or her age of entry into the labor force, contribu- tion density, and decisions regarding the timing of retirement. To some degree, these three factors are influenced by the incentives embodied in the pension system.The tax and contribution system influences net replace- ment rates through the progressiveness of the income tax formula, which taxes (higher) income during a worker's active life more so than it does (lower) pension benefits in retirement. In addition, social security levies (for pensions; unemployment; health care; and, at times, housing and family benefits) are typically reduced or eliminated altogether in retirement.These benefits are particularly important for low- to middle-income groups. The adequacy of income replacement provided by the first-pillar earnings-related pension scheme in the Slovak Republic cannot be 254 Adequacy of Retirement Income after Pension Reforms evaluated without first establishing benchmarks. Unfortunately, there is no consensus on what constitutes adequacy. According to one widely respected definition, pensions are adequate when they are sufficient to prevent poverty among the elderly and provide the vast majority of the population with a reliable mechanism for smoothing income over their lifetime. Even with the benefit of a definition, however, actually establish- ing benchmarks is problematic, because attitudes vary across countries as a result of social and cultural perceptions. Moreover, benchmarks ignore the existence of other factors that affect the welfare of the elderly--and that vary from country to country--including the existence and generosity of health insurance and long-term care, the cost of housing, the structure of traditional living arrangements, the presence of informal intrafamily or intergenerational sources of financial and nonfinancial support, and the availability and security of other mechanisms for people to save for their own retirement. One reputable nine-country study (OECD 2001) observes that liv- ing standards are roughly comparable for people 10 years older than the normal retirement age and people 15 years younger than the normal retirement age when retirees have disposable income equal to roughly 80 percent of the disposable income of working-age people. In part, this is attributable to the fact that retirees have no work-related expenses (they do not have to commute or buy special clothing or uniforms, for example). This finding, however, does not imply that mandatory first- pillar pension schemes should actually target an 80 percent net replace- ment rate. To the contrary, in middle- and high-income countries, one can reasonably expect individuals to save for their own retirement--and the empirical evidence suggests that, in practice, they do so.5 There is also some evidence to suggest that the ratio between pre- and postre- tirement income is somewhat independent of the income replacement mandate of the public pension system. Put simply, individuals tend to save more in countries with smaller mandates (and vice versa). Because the Slovak Republic has access to relatively well-developed financial markets and investing outside the country is permitted with only modest restriction under the third-pillar voluntary pension scheme, it would seem reasonable to expect middle- and higher-income workers to save enough to finance at least 25 percent, if not closer to 50 percent, of this 80 percent income replacement target. Given this, three bench- marks are provided: a 40 percent net replacement rate (which implies that individuals would be expected to save enough to finance half of the total income replacement target); a 60 percent net replacement rate (which implies that individuals would be expected to finance a quarter The Slovak Republic 255 of the target); and an 80 percent net replacement rate (which implies that individuals, most of whom would be low-income earners, would not be expected to contribute anything toward the target).6 In the following analysis, these three benchmarks are used to evaluate the adequacy of benefits in the Slovak Republic compared with the average net replace- ment rate observed in 53 countries around the world, the average net replacement rate observed for selected countries in Europe and Central Asia, and the poverty line in the Slovak Republic. To estimate gross and net replacement rates, we consider two critical dimensions--earnings levels and contribution periods--with the help of the Analysis of Pension Entitlements across Countries (APEX) model.7 This model generates estimates for replacement rates under steady-state assumptions (that is, as if the rules of the reformed pension scheme had been in place over the entire active life of the individual). Because life expectancies at retirement are projected to increase over time--which will affect the benefits paid by defined-contribution pension schemes-- a reference year must be chosen. For this study, 2040 is used, because it provides a sufficiently long contribution period over which to approxi- mate steady-state conditions. The first critical task is to investigate levels of income replacement across a relevant spectrum of income. Income is represented as a percent- age (50­200 percent) of average earnings. The second task is to investigate the impact on income replacement of differences in the duration, timing, and density of an individual's contribution history (density refers to the percentage of time an individual actually contributes over a given period). To facilitate the presentation of these multidimensional results, we com- pute replacement rates as a function of the age at which an individual exits the labor market. They are presented separately for full-career and partial- career workers. Replacement rates for full-career workers. Projected replacement rates for full-career workers in 2040 are examined first. For the purpose of this analysis, a full career is defined as continuous employment from age 20 to the current normal retirement age of 62. Gross replacement rates clearly show why the earnings-related pension scheme has been described as providing a strong link between benefits and contributions (figure 8.3). Gross replacements rates show a 56.7 percent gross replacement rate-- 24.4 percentage points are provided by the points system, and 32.3 percentage points come from the defined-contribution scheme. The situation changes when taxes are taken into consideration (figure 8.4). The impact of taxes and contributions on net replacement 256 Adequacy of Retirement Income after Pension Reforms Figure 8.3 Sources of Gross Replacement Rates in the Slovak Republic, by Income Level 65 55 45 (percent) rate 35 25 replacement 15 gross 5 ­5 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (points system) Source: APEX model. Note: Figure shows projected replacement rate for 2040 as an approximation of steady-state conditions. Figure 8.4 Sources of Net Replacement Rates in the Slovak Republic, by Income Level 100 90 80 (percent) 70 60 rate 50 40 30 replacement 20 net 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage second pillar (defined contribution) first pillar (points system) taxes Source: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. The Slovak Republic 257 rates increases as incomes rise. As a result, high-income workers receive higher net replacement rates than do low-income workers. This is attrib- utable to the progressive nature of the tax code in the Slovak Republic, not to the design of the pension system. Examination of net replacement rates by income level suggests that pensions for most full-career workers in the Slovak Republic can be con- sidered adequate for middle- and high-income workers (figure 8.5).8 Replacement rates for all levels of preretirement income are higher than the middle benchmark, suggesting that the pension system is effectively smoothing consumption from work into retirement for middle- and high- income workers. Replacement rates for low-income workers are lower than the 80 percent benchmark. Moreover, given that benefits for even very low­income full-career workers exceed the poverty line (except, marginally, at the lowest income levels), the objective of poverty allevia- tion is being met. Levels of income replacement for high-income workers in the Slovak Republic are higher than regional and world averages--and Figure 8.5 Net Replacement Rates for Male Full-Career Workers in the Slovak Republic, Europe and Central Asia, and the World 90 high benchmark 80 70 (percent) 60 middle benchmark rate 50 40 low benchmark 30 replacement 20 net 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage world average Europe and Central Asia average Slovak Republic average poverty line (percentage of earnings) Source: Authors' calculations based on World Bank 2005, World Bank 2007a, and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. 258 Adequacy of Retirement Income after Pension Reforms replacement rates for low-income workers are lower than regional and world averages. These findings highlight the strong link between contributions and benefits and illuminate the fact that the Slovak pension system provides relatively little redistribution from the comparatively well-off population to those with lower levels of preretirement income. Replacement rates for partial-career workers. Not everyone works from age 20 to the statutory retirement age. Many individuals enter and exit the labor force (often at different ages and for different periods of time) and earn different wages while working (figure 8.6.). To examine the adequacy of benefits for partial-career workers, we consider three styl- ized cases. (Only middle-income partial-career workers are examined because replacement rates are roughly comparable for workers with lower or higher levels of preretirement income.) These cases include career type A (someone entering the labor force at age 25 who works continu- ously for a period of years before leaving the workforce at some point between the ages of 50 and 70 and then claims a benefit); career type B (identical to career type A, except that the worker enters the workforce at age 30 and leaves no earlier than age 55); and career type C (identical Figure 8.6 Net Replacement Rates for Male Middle-Income Partial-Career Workers in the Slovak Republic, by Career Type and Exit Age 120 100 high benchmark 80 (percent) world and Europe and Central Asia average middle benchmark 60 rate low benchmark 40 poverty line (percentage of earnings) 20 replacement 0 45 50 55 60 65 70 exit age from labor market career type A: career type B: career type C: entry age--25, entry age--30, entry age--25, contribution contribution contribution density--100% density--100% density--75% Source: Authors' calculations based on the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. See text for descriptions of career types. The Slovak Republic 259 to career type A, except that the individual contributes in only three years out of four while in the labor force). In cases in which the withdrawal from the formal labor market occurs before the statutory retirement age, the pension is claimed (and the replacement rate calculated) only at the later age. For withdrawals after the statutory retirement age, the ages coincide. Several conclusions can be drawn from figure 8.6. First, some workers are at risk of receiving levels of income replacement below the poverty line.9 Second, leaving the workforce very early can be very costly. Someone retiring long before reaching the retirement age may not receive levels of income replacement higher than even the lowest of the three benchmarks. Third, entering the workforce later in life is costly. Someone entering the workforce at age 30 receives a net replacement rate that is 6­15 percent- age points lower than someone entering the workforce at age 25. Fourth, working intermittently is costly. Someone who enters the workforce at the same age but who contributes only three years out of four will receive a net replacement rate that is 6­30 percentage points lower than someone who contributes continuously. In all cases, net replacement rates grow faster the longer someone continues to work. This is encouraging, because it demonstrates that the pension system provides incentives for people to remain in the work- force. While career type A and B workers can attain the 40 percent benchmark before reaching the normal retirement age, career type C workers must work until the normal retirement age in order to do so. To replace 60 percent (or 80 percent) of their preretirement earnings, career type A and B workers must work for as many as eight years past the normal retirement age. Career type C workers cannot attain the 80 percent benchmark even if they work until age 70. This does not imply that the pension system is failing to achieve its poverty alleviation objective, however, because the Slovak pension system is supported by a program of social assistance that guarantees all individuals, including the elderly, income equal to the subsistence level. Fiscal Sustainability The sustainability of a pay-as-you-go first-pillar pension scheme is best evaluated in actuarial terms by estimating the scheme's actuarial deficit as the difference between its assets and liabilities. If a large actuarial deficit exists, the scheme is financially unsustainable and needs policy actions that increase its assets, reduce its liabilities, or both. A good proxy for the actuarial deficit is the difference between the present value of the scheme's expected future revenues (that is, contributions and other income) and the expected future expenditures (that is, benefit payments, 260 Adequacy of Retirement Income after Pension Reforms administrative costs, and other expenses) over an extended projection period. The difference between these two values represents an unfunded liability (sometimes referred to as a financing gap) on the public-sector balance sheet. Because this study is also concerned with the time path of revenues and expenditures (and the resulting balance across the projection period ending in 2050), this more pragmatic approach has been taken. Projections of expenditures, revenues, and deficits are presented on the basis of available postreform fiscal projections. Despite the huge improvement in the fiscal condition of the pension system attributable to the parametric reforms of 2004 and the introduc- tion of the second pillar in 2005, the system remains in deficit (figure 8.7). The projected deficit is caused partly by the introduction of the second pillar and the loss of contribution revenues in the first pillar. Revenues are projected to hover around 6 percent of GDP for the entire projection period. Expenditures are projected to fall slightly between 2005 and 2015 but to rise thereafter, albeit at a rate far slower than was projected before the reforms. The growth in expenditures will be driven by the aging of the Slovak population. As a result of the imbalance between revenues and expenditures, deficits are projected to eventually reach levels equivalent to 4.4 percent of GDP by 2050. What options exist for restoring the system to fiscal balance? Unfortunately, for policy makers, the options are limited. Revenues can be Figure 8.7 Projected Fiscal Balance of the Slovak Republic's Public Pension Scheme after Reform, 2005­50 12 10 8 6 GDP of 4 2 0 percentage ­2 ­4 ­6 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 expenditures revenues deficit Source: Unpublished World Bank Pension Reform Options Simulation Toolkit (PROST) simulations. The Slovak Republic 261 increased by increasing the contribution rate and using general revenues to compensate partly or fully for the transition costs toward the funded second pillar. Alternatively--or in addition, because the options are not mutually exclusive--expenditures can be reduced by cutting benefits, increasing the minimum number of years required to become eligible for benefits, or delaying the payment of benefits by raising the retirement age further. Because raising the contribution rate could threaten competitive- ness and will likely strengthen incentives for tax evasion, it is typically not embraced by policy makers. This leaves policy makers with limited options: cutting benefits, tightening eligibility conditions, or raising the retirement age. It also raises the question of whether restoring sustainabil- ity will exact a cost in terms of the adequacy of benefits provided to future beneficiaries. Increasing the retirement age further (commensurate with expected changes in life expectancy) would reduce the long-term deficit to 2.9 percent of GDP. Fully restoring long-term fiscal balance would require that retirement ages be gradually increased to 74 by 2050 (unpublished World Bank PROST simulations). If retirement ages are left unchanged and the current structure of the system is retained, further cuts in benefits--on the order of a 36 percent reduction in the average benefit provided under the first pillar--will be required if the system is to become sustainable. Restoring sustainability will exact a cost in terms of the adequacy of benefits provided to future beneficiaries: if benefits are adjusted to maintain a fiscal balance similar to the current level in proportion to the overall size of the first-pillar scheme, full-career workers will receive replacement rates roughly 9 per- centage points lower in 2050 than they receive today (figure 8.8). Two observations emerge from a comparison of these new (and lower) net replacement rates against the three benchmarks. First, a 36 percent reduction in first-pillar benefits would not result in income replacement for full-career workers falling below the poverty line, except for very-low- income workers. This indicates that a sustainable first-pillar pension scheme in the Slovak Republic would still achieve its poverty alleviation objective. Second, a 36 percent reduction in benefits would still support the objective of smoothing lifetime consumption for middle- and high- income full-career workers, because levels of income replacement are still equal to or higher than the 60 percent benchmark. Low-income workers, however, would fall even further below the 80 percent benchmark. This observation is subject to three caveats. First, this analysis consid- ers only full-career workers, while the average worker now contributes for only about 27­30 years, substantially less than the 40 years expected of a 262 Adequacy of Retirement Income after Pension Reforms Figure 8.8 Net Male Replacement Rates in the Slovak Republic before and after Benefit Adjustment 100 80 high benchmark (percent) 60 rate middle benchmark 40 low benchmark replacement 20 net 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage net replacement rate adjusted net replacement rate poverty line (percentage of average income) Source: Authors' calculations based on the APEX model. full career'. Contributing to the pension scheme for only 30 years, for example, reduces net income replacement by 18 percentage points (for low-income workers) and by 22 percentage points (for high-income work- ers). Second, if benefit cuts are combined with further increases in the retirement age, the benefit cuts will not need to be as steep in order to restore fiscal balance. Third, workers always have the option of saving out- side the first-pillar pension scheme. To increase income replacement by one percentage point, for example, a full-career worker would need to save only about 0.5 percent of his or her earnings from age 40 to the cur- rent age of retirement.10 Conclusions In response to a looming crisis in the existing pay-as-you-go public pen- sion system and the projections that suggested that the situation would gradually worsen as a result of the aging population, the Slovak Republic introduced significant parametric reforms to its first-pillar pension scheme in 2004. As a means of increasing revenues, retirement ages were increased to extend the period over which workers contribute to the system and the mechanism by which benefits are computed was The Slovak Republic 263 changed to strengthen the link between preretirement contributions and postretirement benefits, thereby improving compliance incentives. So the growth in expenditure could be curtailed, average benefits were reduced. As a result of these reforms, the long-term fiscal position of the pension system improved substantially, with projected deficits for 2050 falling from 10.0 percent of GDP to 4.4 percent of GDP. In 2005, the Slovak Republic introduced a mandatory second pillar to provide workers with a mechanism for diversifying their retirement income. Together with the first-pillar reforms, the introduction of the sec- ond pillar eliminated the highly redistributive provisions of the old pen- sion system that had created incentives for evasion and the underreporting of income and strengthened incentives for people to remain in the work- force after reaching the minimum retirement age. The reform, however, also contributed to an increase in the cash deficit, because contribution revenues from the first pillar were diverted to second-pillar financing. The resulting gross and net replacement rates for full-career workers are projected to be in line with regional and world averages (see chapter 1).The future net replacement rates for full-career workers are projected to be about 70­80 percent across the analyzed income spectrum. As in other countries, workers with less than full careers--because they left the work- force before reaching retirement age, worked intermittently, or have gaps in their employment history--risk receiving a level of income replacement closer to the lower benchmark of 40 percent, or possibly even lower. To restore the long-term fiscal balance of the first-pillar pension scheme without recourse to general revenue financing, policy makers could raise retirement ages--rough estimates suggest to 74 by 2050. Realizing the full fiscal impact of this measure requires maintaining income replacement at levels now associated with the current retirement age. Increasing the retirement age in step with increases in life expectancy at retirement is a natural choice, for both individuals and policy makers, but it requires cross-sectoral policy reforms that would enable elderly workers to con- tinue to participate in the labor market.11 Other options for restoring fiscal balance include cutting benefits at retirement, reducing the generosity of benefit indexation, or adopting some combination of the two. Rough estimates suggest that average replacement rates at retirement would have to fall by some 9 percentage points. Moving from Swiss indexation to full price indexation would reduce average first-pillar benefits by some 10 percentage points. Reducing benefits by this amount would not compromise the objective of smooth- ing lifetime consumption for middle- and high-income full-career workers 264 Adequacy of Retirement Income after Pension Reforms or alleviating poverty among the elderly, because levels of income replace- ment would still not fall below the poverty line. Levels of income replace- ment for low-income individuals, however, would fall below the high benchmark following such a reduction. Individuals who wish to defer more of their lifetime consumption into retirement would still have the option, of course, of participating in the voluntary third-pillar pension scheme. Notes 1. The number of contributors declined 14 percent between 1995 and 2002 (World Bank 2004). 2. A notional scheme computes benefits using a traditional defined-contribution formula (where the rate of return on the notional balance is computed using an economic proxy, such as the rate of growth in economywide wages) but finances those benefits using traditional pay-as-you-go financing. This approach tightens the connection between contributions and benefits with- out exposing participants to investment risks or imposing transition costs (see Holzmann and Palmer 2006). 3. Under the prereform rules, only the first Sk 2,500 was fully counted toward the pension assessment base. For earnings of Sk 2,501­SK 6,000, only a third were counted; for earnings of Sk 6,001­Sk 10,000, only a tenth were counted; earnings of more than Sk 10,000 were excluded. As a result, contributions on earnings of more than SK 10,000 were pure taxes (World Bank 2004). 4. Legislation states that the second pillar is not mandatory for new entrants.These projections assume that all new entrants elect to participate in the second pillar. 5. In Chile, for instance, 70 percent of retirees from the mandatory public pension system own their home, which is a form of savings (see Valdés-Prieto 2008). 6. These benchmarks approximate the standards developed by the International Labour Organization (ILO) (1952) and the Council of Europe (1990). ILO Convention 102 of 1952 sets a minimum benefit equal to 40 percent of the reference wage for married men of pensionable age. This amount was raised to 45 percent in 1968. The European Code of Security of 1990 sets a mini- mum standard for members of the Council of Europe equal to 65 percent for married people of a specific age. 7. The APEX model was developed by Axia Economics, with funding from the Organisation for Economic Co-operation and Development and the World Bank. The model codes detailed eligibility and benefit rules for first- and second-pillar schemes by using available public information that has been verified by country contacts. Because the details of the rules sometimes change on short notice (and limited public disclosure), the calculations pre- sented here should be considered as best approximations only. The Slovak Republic 265 8. Replacement rates are simulated for an unmarried man working a hypothetical career path under the assumption that real wage growth is 2 percent, inflation is 2.5 percent, the rate of return on invested assets is 3.5 percent, and the worker retires at the statutory retirement age. Replacement rates shown do not consider the benefits received from occupational schemes. 9. A figure of 35 percent of the average net wage is used as a proxy for the poverty line, because this percentage very broadly approximates a US$2.25- a-day poverty line converted into national currency, adjusted for purchasing power parity, expressed relative to the national average net wage, and aver- aged across the eight study countries. 10. This estimate is based on the assumption that real wage growth is 2 percent, the net real rate of return on invested assets is 3.5 percent, and benefits (from both the unfunded and the funded pillars) are price indexed. 11. See Holzmann, MacKellar, and Repansek (2009) for a conference volume that addresses theses issues for the countries of southeastern Europe. Bibliography Cillíková, J. 2006. "Trends and Reforms in Slovakia." Paper presented at the CEE (Central and Eastern Europe) forum, Vienna, November 23. www.gvfw.at/ files/slovakia.pdf. Council of Europe. 1990. European Code of Social Security (Revised). Rome. European Commission. 2007a. Mutual Information System and Social Protection (MISSOC) database. ec.europa.eu/employment_social/. ------. 2007b. "Pension Schemes and Projection Models in EU-25 Member Countries." European Economy Occasional Paper 37, Economic Policy Committee and Directorate General for Economic and Financial Affairs, Brussels. GVG (Gesellschaft für Versicherungswissenschaft und -gestaltung) and European Commission. 2003. Study on the Social Protection Systems in the 13 Applicant Countries: Slovak Republic Country Report. Cologne. Hlavacka, S., R. Wagner, and A. Rosenberg. 2004. Health Care Systems in Transition: Slovakia. Geneva: World Health Organization. Holzmann, R., and R. Hinz. 2005. Old-Age Income Support in the 21st Century. Washington, DC: World Bank. Holzmann, R., L. MacKellar, and J. Repansek, eds. 2009. Pension Reform in South- eastern Europe: Linking to Labor and Financial Market Reforms. Washington, DC: World Bank. Holzmann, R., and E. Palmer, eds. 2006. Pension Reform: Issues and Prospects for Non-Financial Defined Contribution Schemes. Washington, DC: World Bank. 266 Adequacy of Retirement Income after Pension Reforms ILO (International Labour Organisation). 1952. ILO Convention 102. Geneva: ILO. ------. 1967. ILO Convention 128. Geneva: ILO. INPRS (International Network of Pension Regulators and Supervisors). 2003. Complementary and Private Pensions throughout the World. Geneva: INPRS. Institute of Financial Policy. 2006. The Pension System in the Slovak Republic. Bratislava. Lesay, I. 2006. Pension Reform in Slovakia: The Context of Economic Globalization. Brussels: European Trade Union Institute. Lindbeck, A., and M. Persson. 2003. "The Gains from Pension Reform." Journal of Economic Literature 41 (March): 74­112. Ministry of Labor, Social Affairs and Family. 2005. National Strategy Report on Adequate and Sustainable Pensions. Bratislava. Nalevanko, M. 2005. Pension Reform in Slovakia: Why and How? Bratislava: Ministry of Labor, Social Affairs and Family. Natali, D. 2004. The Reformed Pension System. Observatoire Social Européen. OECD (Organisation for Economic Co-operation and Development). 2001. Ageing and Income: Financial Resources and Retirement in Nine OECD Countries. Paris: OECD. Queisser, M., and E. Whitehouse. 2006. Neutral or Fair? Actuarial Concepts and Pension System Design. Paris: Organisation for Economic Co-operation and Development. Reiterer, A. 2008. Population Development and Age Structure in Southeastern Europe until 2050. World Bank, Washington, DC. U.S. Social Security Administration. 2006. Social Security Systems throughout the World: Europe. Washington, DC: Social Security Administration. Valdés-Prieto, S. 2008. Designs for the First-Pillar Pensions and the 2008 Chilean Reform. http://editorialexpress.com/cgi-bin/conference/download.cgi?db_ name=SECHI2008&paper_id=130. Whitehouse, E. 1999. Tax Treatment of Funded Pensions. Washington, DC: World Bank. World Bank. 2002. Slovak Development Policy Review.Washington, DC:World Bank. ------. 2004. Slovak Republic: Pension Policy Reform Note. Washington, DC: World Bank. ------. 2005. The Quest for Equitable Growth in the Slovak Republic. Washington, DC: World Bank. ------. 2007a. Pensions Panorama. Washington, DC: World Bank. ------. 2007b. Social Assistance in Central Europe and the Baltic States. Washington, DC: World Bank. WHO (World Health Organization). 2008. Database. http://www.who.int/ research/en/. C H A P T E R 9 Slovenia Slovenia inherited from the former Yugoslavia a traditional pension system financed on a pay-as-you-go basis (meaning that contributions from current workers are used to pay benefits to current beneficiaries). Following the transition from a state-planned economy to a market economy in 1991, the pension system began generating deficits, which, in 1992, were equivalent to 0.3 percent of gross domestic product (GDP). Pension expenditure increased as the number of retirees grew. In 1996, the government reduced the contribution rate, which led to a sharp drop in contribution revenues. By 1996, the fiscal deficit had reached 3.5 percent of GDP. Deficits were projected to grow further as the population ages. The old-age dependency ratio (the population age 65 and older divided by the population age 20­64) was expected to increase from 23.5 percent in 2005 to 60.3 percent by 2050 (see Reiterer 2008). Recognizing these challenges, Slovenian policy makers introduced sev- eral parametric changes to the existing public pension system. Rather than adopting a multipillar approach, as many other transition economies did, in 2000 Slovenia redesigned its existing pay-as-you-go pension system. Recognizing that benefits are expected to decrease following the reform of the public scheme, it also introduced a voluntary fully funded, defined- contribution scheme to provide individuals with additional savings options. 267 268 Adequacy of Retirement Income after Pension Reforms Despite these reforms, the aging of the population will continue to place stress on Slovenia's pension system. Projections suggest that deficits will reach 8.8 percent of GDP by 2050, which will almost certainly com- pel the government to introduce further reforms. Against this backdrop, this chapter evaluates the Slovenian pension system, focusing on fiscal sustainability and benefit adequacy. Adequacy is evaluated through the lens of statutory net replacement rates for differ- ent retirement ages, patterns of contributions, and income levels with comparisons to international benchmarks. This chapter is organized as follows. The next section discusses the motivation for the reforms. The following section describes the key char- acteristics of the reformed pension system. The third section assesses the adequacy of pension benefits and the fiscal sustainability of the system. The last section draws conclusions. Motivation for Reform Upon gaining independence in 1991, Slovenia inherited a socialist-era public pension system financed on a pay-as-you-go basis. The system suf- fered from a number of serious design flaws similar to those observed in other transition economies, including privileges for certain occupations, low retirement ages, generous benefits, and loose eligibility conditions. These characteristics affected both the revenues and the expenditures of the system, leading to fiscal imbalances beginning in 1992, when the sys- tem generated a deficit (table 9.1). To address these issues, in 1992 Slovenia passed a reform law that included gradually raising the retirement age and restricting the criteria for early retirement. These changes came too late to prevent a large Table 9.1 Fiscal Balance of Slovenia's Pension System before Reform, 1992­99 (percentage of GDP) Year Revenues Expenditures Balance 1992 13.5 13.8 ­0.3 1993 13.9 14.4 ­0.5 1994 13.1 14.4 ­1.3 1995 12.9 14.7 ­1.8 1996 11.0 14.5 ­3.5 1997 10.1 14.4 ­4.3 1998 10.2 14.3 ­4.1 1999 9.9 14.4 ­4.5 Source: ILO 2004. Slovenia 269 number of early retirees from joining the benefit rolls from 1990 to early 1992.1 Following a cut of 6.65 percentage points in the contribu- tion rate paid by employers, the fiscal balance of the pension system deteriorated further, reaching 3.5 percent of GDP by 1996. By 1999, the deficit had reached 4.5 percent of GDP, despite the fact that contribu- tion rates remained constant after 1996. In the absence of further reforms, expenditure was projected to eventually reach 26 percent of GDP (World Bank and IIASA 2001). Recognizing these challenges, in 1999 the government introduced additional parametric reforms. These reforms included tightening the eligibility criteria for benefits, reducing accrual rates, and increasing the number of years of service required to collect benefits. Characteristics of Slovenia's Pension System This section describes the main characteristics of Slovenia's pension system. These include the design of the individual pillars of social insurance; the rules governing pension system taxation, institutional structure, and cover- age; and the provisions governing old-age, disability, and survivorship pen- sions. The design of the pension system is assessed using a conceptual framework developed by the World Bank, which generally recommends including a funded component if conditions are appropriate but increas- ingly recognizes that a range of choices is available to policy makers to pro- vide effective old-age protection in a manner that is fiscally responsible (see Holzmann and Hinz 2005). In general, the World Bank supports pension systems composed of some combination of five basic pillars: · a noncontributory (or zero) pillar (in the form of a demogrant, social pension, or social assistance benefit) intended to provide a minimal level of income protection; · a first-pillar contributory system linked to earnings, which seeks to replace a portion of preretirement income; · a mandatory second pillar (essentially, individual savings accounts), which can be designed in various ways; · a voluntary third pillar, which is flexible and discretionary (this pillar, too, can take a variety of forms); and · a fourth pillar of informal intrafamily or intergenerational sources of financial and nonfinancial support to the elderly, including access to health care and housing. 270 Adequacy of Retirement Income after Pension Reforms Pillar Design The design of the Slovenian pension system incorporates four of the five pillars recommended by the World Bank (table 9.2). The publicly managed noncontributory zero pillar, financed with general tax rev- enues, redistributes income to lower income groups using means test- ing, such that eligible beneficiaries receive a benefit sufficient to ensure them of a total income equal to a third of the state-defined minimum pension assessment base. The mandatory first pillar is an earnings- related, defined-benefit pension scheme financed on a pay-as-you-go basis. There is no national second-pillar pension scheme.2 The voluntary third pillar, introduced in 2000 to supplement the benefits of the mandatory first pillar, is a defined-contribution scheme in which bene- fits depend on an individual's contributions and investment earnings at the point of retirement. The mandatory fourth pillar, financed by a combination of contributions and copayments, provides health insur- ance to all people, including the elderly. Benefits paid by the noncontributory zero pillar are exempt from tax- ation. Contributions to the first pillar are exempt from taxation, while benefits are taxed. The third pillar is subjected to a exempt-exempt-taxed (EET) regime, meaning that contributions are partially exempt from tax- ation, investment income is fully exempt, and benefits are taxed (see box 1.1 in chapter 1).This is similar to the tax regimes of most of the coun- tries of the Organisation for Economic Co-operation and Development (OECD), which also take an EET approach. Contributions to the fourth pillar (the health care system) are exempt from taxation. Noncontributory scheme. Individuals over age 65 who have lived in Slovenia for at least 30 years between the ages of 15 and 65 and who do not qualify for a pension from the first-pillar scheme are eligible for a means-tested noncontributory pension equal to a third of the minimum pension assessment base (European Commission 2007a). In 2006, such benefits were paid to 3.3 percent of people over age 65, at a cost equiv- alent to 0.3 percent of GDP. Individuals who qualify for a pension from the first-pillar scheme but who receive very low benefits may apply for a pension income supple- ment, provided that their pension is lower than the minimum pension for a full contribution period and their household does not have other sources of income sufficient to exceed the minimum living standard. The supplement is calculated by multiplying a coefficient (which varies pro- portionally in relation to the length of the individual's contribution period) Table 9.2 Structure of Slovenia's Pension System Taxation Investment Generic Benefit income/capital Scheme type Coverage Type Function Financing benefit indexation Contributions gains Benefits Zero pillar Universal Means tested Redistributive Tax revenues 33.3 percent Growth of n.a. n.a. Exempt (public non- of minimum minimum contributory) pension pension assessment assessment base base First pillar Mandatory Defined Insurance Percentage of Benefit Wage Exempt n.a. Taxed (public, earn- benefit individual calculated on growth ings related) earnings the basis of pension assessment base and accrual rate Third pillar Voluntary Defined Insurance Voluntary Pension Depends Exempta Exempt Taxed (private, earn- contribution contributions from capital on options ings related) accumulation chosen Fourth pillar Mandatory n.a. Insurance Percentage of Specified n.a. Exempt n.a. n.a. (public individual basic health health care) earnings service package plus tax revenues Sources: European Commission 2007a, 2007b. n.a. = Not applicable. 271 a. An amount up to 24 percent of contributions to the first pillar is tax exempt. 272 Adequacy of Retirement Income after Pension Reforms by the difference between the minimum pension for a full contribution period and the actual pension the individual receives. Earnings-related schemes. Slovenia's traditional mandatory single-pillar earnings-related, pay-as-you-go, public pension scheme provides old-age, disability, and survivor benefits.3 One of the major reforms of 2000 was to change the benefit formula (table 9.3). Before reform, the minimum level of income replacement for 15 years of service was 35 percent for men and 40 percent for women. The accrual rate for each additional year of service was 2 percent, with an income replacement ceiling of 85 percent. Women were awarded 3 percent per Table 9.3 Parameters of First-Pillar Earnings-Related Scheme in Slovenia before and after Reform Pension Vesting Contribution Contribution assessment Retirement Period period rate ceiling Benefit rate base age Prereform 15 years 31 percent No maximum 40 percent 10 best 58 for men, (15.5 percent for men and consecu- 53 for by employer, 35 percent tive years women 15.5 percent for women by employee) for first 15 years, 2 percent thereafter Postreform 15 years 24.35 percent No maximum 35 percent Gradually Gradually (15.50 for men and increased increased percent by 38 percent to reach to reach 63 employee, for women best 18 for men in 8.85 percent for first 15 years in 2009 and by employer) years, 1.5 2008 61 for percent per Wages are women year beyond valorized in 2023 15 years using coef- ficients determined from growth of wages and pensions in preceding years Sources: European Commission 2007a, 2007b; World Bank and IIASA 2001. Slovenia 273 year up to 20 years of service. As a result, women could qualify for the maximum replacement rate with 35 years of service while men needed 40 years. The pension assessment base was calculated based on the 10 best consecutive years of service. The reforms changed this such that accrual rates for both men and women are now 1.5 percent per year. The minimum replacement rate for 15 years of service is maintained at 35 percent for men but is reduced to 38 percent for women, equalizing the replacement rate for a full-service career (defined as 40 years for men and 38 years women) at 72.5 percent. For ease of transition, contribution periods before 1999 are still evaluated using prereform provisions. Before the reforms of 2000, the pension assessment base was calcu- lated using an individual's 10 best years of earnings. The reforms gradu- ally changed this to the 18 best years of earnings, effective as of 2008. Wages are valorized using coefficients determined from the growth of wages and pensions in the preceding years. The value of the revaluation coefficient is established yearly by ministerial decree; depending on the year, values are 77­80 percent of the growth of nominal wages.4 A mini- mum and a maximum are applied to the pension assessment base. If the actual pension assessment base is lower than the statutory minimum assessment base, the pension is calculated using the statutory minimum.5 The maximum pension assessment base is four times the minimum. The reforms also raised retirement ages. The retirement age for men was increased by six months per year to reach 63 by 2009. The retirement age for women was increased by four months per year to reach 61 by 2023. The years of service required to become eligible for a full pension were also gradually increased, to 40 for men and 38 for women. Postretirement pensions are indexed to wages, with indexation conducted twice a year. Benefits are financed by contributions from employers and employees. The contribution rate is 15.50 percent for employers and 8.85 percent for employees, for a total levy of 24.35 percent. There is no ceiling on wages for contributions. Voluntary scheme. The voluntary, defined-contribution, third-pillar pension scheme was introduced in 2001 to supplement the benefits provided under the mandatory first pillar (table 9.4). To be eligible for benefits, workers must have claimed their old-age pension from the first pillar, be at least 58 years old, and have contributed to the volun- tary scheme for a minimum of 120 months. Investments are guaran- teed to earn at least 40 percent of the average annual interest rate paid on fixed-income government securities with a maturity of more than 274 Adequacy of Retirement Income after Pension Reforms Table 9.4 Characteristics of the Voluntary Scheme in Slovenia Lump-sum Tax advan- Contributions payments Vesting Retirement tages to tax deductible possible in Coverage period age participants by employers retirement Individuals covered 10 years 58 Yes Yes Yes under the public scheme Source: European Commission 2007a. one year. The Insurance Act regulates investments and gives pension fund managers wide latitude to purchase securities, make loans, purchase real estate, and make bank deposits. There are restrictions on asset class allocation and foreign investments. The voluntary pension scheme is administered by mutual funds, pension fund companies, and insurance companies authorized to sell life insurance. Mutual funds are licensed by the Securities Market Agency; pension fund and insurance compa- nies are licensed by the Insurance Supervision Agency. Individuals can enroll either individually or collectively through an employer. They may elect to join a fund individually while paying their contributions through an employer. In 2006, 196,883 participants (17 percent of the labor force) were contributing to four pension fund companies. The total assets of the scheme amounted to 22.9 billion tolars (SIT) (0.3 percent of GDP) (Insurance Supervision Agency 2006).6 Health care system. Health care in Slovenia is provided primarily by a mandatory health insurance scheme administered by the Health Insurance Institute of Slovenia. The scheme is financed mainly by contributions from the covered population. The contribution rate for insured employ- ees is 12.92 percent of payroll (6.56 percentage points of which are paid by employers and 6.36 percentage points of which are paid by employ- ees). Pensioners pay 5.65 percent of their gross pension. Self-employed workers contribute 12.92 percent and farmers contribute 6.36 percent of their net income. Individuals with no income are registered in municipal- ities, which are obliged to pay a fixed contribution (Albreht and others 2002; U.S. Social Security Administration 2006). In 2005, health expenditure accounted for 8.5 percent of GDP, 71.9 percent of which was public expenditure and 28.1 percent was private expenditure. Of the private expenditure, 88.1 percent was attrib- utable to out-of-pocket expenditure (informal payments, direct payments, and copayments) (WHO database). Slovenia 275 The benefit package provided by the mandatory health insurance scheme covers almost all services, including preventive services, diagnos- tic procedures, ambulatory care, and long-term nursing care. Almost all services require copayments, which are 5­50 percent of the cost of the service. Individuals may make copayments directly or purchase copay- ment insurance (which is becoming quasi-mandatory). The government recently proposed a bill to Parliament relating to the provision of cover- age for people of low income. Institutional Structure and Coverage of Earnings-Related Schemes The Pensions and Disability Insurance Institute (PDII) administers the first-pillar scheme, using a centralized database and payment system. The PDII also has regional offices, which play a purely administrative role. Legal supervision of the PDII is carried out by the Ministry of Labor, Family and Social Affairs. Contributions and personal income taxes are collected simultaneously by a separate agency. The first pil- lar is mandatory for both employees with an established employer relationship and the self-employed. Voluntary participation is permit- ted for certain categories of people, as defined in the law. In 2006, 857,922 individuals were contributing to the first-pillar scheme (about 61 per- cent of the total working-age population and 83 percent of the labor force) and 536,887 individuals (about 27 percent of the population) were receiving pensions. Structure of Benefits The first-pillar earnings-related pension scheme provides old-age, disabil- ity, and survivorship pensions. The provisions governing each of these types of benefits are discussed as follows. Old-age benefits. To be eligible for an old-age pension, individuals must have at least 15 years of service. The minimum old-age pension is 35 per- cent of the minimum pension assessment base for men and 38 percent for women. For years of service beyond 15, the accrual rate is 1.5 percent per year. Pensions of eligible individuals with very low earnings are com- puted using the statutory-minimum pension assessment base (set at about 64 percent of the average wage in 2003). Pensions of individuals with very high earnings are computed using the maximum pension assessment base (equal to four times the minimum assessment base). Retirement ages are being gradually increased to 63 for men (by 2009) and 61 for women (by 2023). The number of years of service required to 276 Adequacy of Retirement Income after Pension Reforms become eligible for a full pension is also being increased, to 40 years for men and 38 years for women. Once this requirement is met, workers will be able to retire at age 58. Men and women with 20 years of service may retire at 63 and 61, respectively. Men and women with 15 years of serv- ice may retire at 65 and 63, respectively. Both mothers and fathers may retire earlier as a function of the number of children they have.7 The retirement age is reduced by eight months for one child, 20 months for two children, 36 months for three children, and 36 months plus an addi- tional 20 months per child beyond three. The minimum retirement age for women with children is 56. Higher accrual rates are provided to men who reach age 63 and women who reach age 61 and who are eligible for benefits but elect to defer their retirement.8 These rates (which vary by year) are provided for a maximum of four years, after which additional benefits accrue at the rate of 1.5 percent a year. Penalties apply to men age 58­63 who have less than 40 years of contributions and to women age 58­63 who have less than 38 years of contributions. The severity of the penalty varies depending on the age of the individual.9 Disability benefits. Disability benefits in Slovenia are based on the cause of the disability (table 9.5). For occupational diseases or work- related injuries, benefits are paid regardless of the individual's period of contributions. If the disability results from other causes, individuals must have paid contributions for at least a third of the period between age 20 and the date of their disability to be eligible for benefits. Disability benefits are calculated in a manner similar to that used to cal- culate old-age pensions, but benefits cannot be lower than 45 percent Table 9.5 Eligibility Conditions for and Benefits Provided by Disability Pensions in Slovenia under the First-Pillar Earnings-Related Scheme Vesting period Contributions Eligibility Benefit rate Partial pension Individual must No specific At least 30 Based on level Prorated based have contributed contributions percent loss of disability; on level of dis- at least one-third for disability of capacity 10­24 percent ability of the period to work of minimum between age 20 pension for full and the time of pension qualify- disability ing period Source: European Commission 2007a. Slovenia 277 for men and 48 percent for women of the minimum pension assess- ment base. Survivor benefits. Survivor benefits are awarded to the dependents of individuals who at the time of death were receiving (or had met the cri- teria to receive) an old-age or disability pension (table 9.6). The mini- mum eligibility age for widows and widowers is 53. If a widow or widower is incapable of working, becomes incapable of working within a year of the individual's death, or has dependent children, survivor pen- sions are awarded irrespective of age. Survivor benefits are set at at least 45 percent of the deceased's pen- sion assessment base. Widow and widower benefits are 70 percent of the higher of the old-age or disability pension, provided there are no other survivors. If the spousal survivor is already receiving an old-age or disability pension, the maximum replacement rate is 15 percent of the deceased's pension, not to exceed the average monthly pension in Slovenia the pre- vious year. Orphans below age 15 (26 for students) are eligible for orphans' benefits. If the orphan is incapable of working, benefits are paid for life. Dependent parents, grandchildren, and siblings of the deceased are also eligible for benefits under certain conditions. Benefits are provided to dependent parents regardless of age if they are incapable of working. Total survivor benefits paid to all survivors cannot exceed 100 percent of the deceased's pension. Table 9.6 Eligibility Conditions for and Benefits Provided by Survivor Pensions in Slovenia under the First-Pillar Earnings-Related Scheme Spouse Orphan replacement Benefit Remarriage Orphan replace- Total family Eligibility rate duration test age limit ment rate benefit Eligibility of 70 percent of For life if Benefits 15 (26 for 70 percent 100 percent deceased deceased's spouse is cease if students) if sole ben- regardless for old-age pension if 70 percent survivor eficiary of number or disability sole benefici- disabled, is remarries of survivors pension ary; 15 taking care before percent of of a child, reaching deceased's or has retirement pension if reached age unless receiving retirement incapable own age; other- of working pension wise, one year Source: European Commission 2007a. 278 Adequacy of Retirement Income after Pension Reforms Assessment of the Performance of Slovenia's Pension System The World Bank has established four principles for evaluating public pen- sion systems, which together should guide the process of pension reform (see Holzmann and Hinz 2005). Broadly speaking, these principles include the adequacy and security of benefits, the affordability of contri- butions, the sustainability of the system over time, and the robustness of the system in the face of demographic changes and macroeconomic shocks. This chapter focuses primarily on the adequacy of benefits and financial sustainability of the first-pillar, earnings-related pension scheme. The remaining principles are mentioned only briefly. Adequacy is ana- lyzed through the lens of net replacement rates. Financial sustainability is evaluated using projections of pension expenditure and revenues. Benefit Adequacy Replacement rates are a useful yardstick for measuring the adequacy of pension benefits, because they express benefits relative to preretirement earnings, thereby indicating the degree to which income is replaced when workers retire. Two variants are commonly used. Gross replacement rates compute income replacement as the ratio of benefits paid to pretax pre- retirement earnings. Net replacement rates compute income replacement as the ratio of benefits received (that is, after the payment of taxes and other levies, including contributions for social insurance) to posttax pre- retirement earnings. In general, net replacement rates are a more useful measure of benefit adequacy, because they capture the degree to which actual take-home pay is replaced when workers retire. The level of income replacement at retirement is not the only measure of benefit adequacy. To fully assess benefit adequacy, it is also important to determine how postretirement indexation rules will affect replacement rates during retirement. Pension benefits in retirement are expected to be indexed to inflation, so that their real value is maintained. In a growing economy with rising real wages, however, mere price indexation of pen- sions leads to a deterioration of the relative consumption position of the retirees. Individuals with otherwise identical work histories will receive dif- ferent pensions depending on when they retire. For this reason, some countries have introduced mixed indexation of pensions that use varying weights of inflation and wage growth in the indexation formula. For an evaluation of the effect of indexation on replacement rates in Slovenia, the replacement rates are normalized to 100 and the assumptions for calcu- lating the replacement rates are maintained (that is, inflation is 2.5 percent Slovenia 279 per year and real wage growth is 2 percent per year). The change in the replacement rate is measured in comparison to full wage indexation or an active worker. The results of the analysis indicate that the relative income position of retirees will be maintained because pensions are indexed to wage growth in Slovenia. The evaluation of income replace- ment that follows considers replacement rates only at retirement. Given Slovenia's indexation policies, however, replacement rates are preserved in retirement. Replacement rates are a function of the formula governing pension benefits; an individual's contribution history; and, in the case of the net replacement rates, the rules of income tax, social security contributions, and other relevant levies. The benefit formula establishes the degree to which the system redistributes income across individuals of different lev- els of preretirement earnings. Progressive systems provide higher levels of income replacement to people with lower levels of preretirement income. In general, the degree to which a system is redistributive depends on the existence (and value) of flat transfers and minimum pension guarantees, the degree to which benefits are earnings-related, and the existence of ceilings on earnings subject to contributions. An individual's contribution history can be characterized by his or her age of entry into the labor force, contribution density, and decisions regarding the timing of retirement. To some degree, these three factors are influenced by the incentives embod- ied in the pension system. The tax and contribution system influences net replacement rates through the typical progressiveness of the income tax formula, which taxes higher income during a worker's active life more so than it does lower pension benefits in retirement. In addition, social secu- rity levies (for pensions; unemployment; health care; and, at times, hous- ing and family benefits) are typically reduced or eliminated altogether in retirement. These benefits are particularly important for low- to middle- income groups. The adequacy of income replacement provided by the first-pillar earnings-related pension scheme in Slovenia cannot be evaluated with- out first establishing benchmarks. Unfortunately, there is no consensus for what constitutes adequacy. According to one widely respected defi- nition, pensions are adequate when they are sufficient to prevent poverty among the elderly and provide the vast majority of the population with a reliable mechanism for smoothing income over their lifetime. Even with the benefit of a definition, however, establishing benchmarks is problematic, because attitudes vary from one country to another as a function of social and cultural perceptions. Moreover, benchmarks ignore 280 Adequacy of Retirement Income after Pension Reforms the existence of other factors that affect the welfare of the elderly--and that vary from country to country--including the existence and gen- erosity of health insurance and long-term care, the cost of housing, the structure of traditional living arrangements, the presence of informal intrafamily or intergenerational sources of financial and nonfinancial support, and the availability and security of other mechanisms for saving for one's own retirement. One reputable nine-country study (OECD 2001) observes that living standards are roughly comparable for people 10 years older than the normal retirement age and people 15 years younger than the normal retirement age when retirees have disposable income equal to roughly 80 percent of the disposable income of working-age people. In part, this is attributable to the fact that retirees have no work-related expenses (they do not have to commute or buy special clothing or uniforms, for exam- ple). This finding, however, does not imply that mandatory first-pillar pension schemes should actually target an 80 percent net replacement rate. To the contrary, in middle- and high-income countries, one can rea- sonably expect individuals to save for their own retirement--and the empirical evidence suggests that in practice they do so.10 There is also some evidence to suggest that the ratio between pre- and postretirement income is somewhat independent of the income replacement mandate of the public pension system. Put simply, individuals tend to save more in countries with more modest mandates (and vice versa). Because Slovenia enjoys a relatively well-developed banking sector, has access to established financial markets, and permits external investments under the third pillar with only modest restrictions, it would seem reason- able to expect middle- and higher-income workers to save enough to finance at least 25 percent, if not closer to 50 percent, of this 80 percent income replacement target. Given this, three benchmarks are provided: a 40 percent net replacement rate (by implication, individuals would be expected to save enough to finance half of the total income replacement target), a 60 percent net replacement rate (individuals would be expected to finance a quarter of the target), and an 80 percent net replacement rate (individuals, most of whom would be low income earners, would not be expected to contribute anything toward the target).11 In the following analysis, these three benchmarks are used to evaluate the adequacy of benefits in Slovenia compared with the average net replacement rate observed in 53 countries around the world, the average net replacement rate observed for selected countries in Europe and Central Asia, and the poverty line in Slovenia.12 Slovenia 281 To estimate gross and net replacement rates, we use the Analysis of Entitlements across Countries (APEX) model to consider two critical dimensions: earnings levels and contribution periods.13 This model gen- erates estimates for replacement rates under steady-state assumptions (that is, as if the rules of the reformed pension scheme had been in place over the whole active life of the individual). Because life expectancies at retirement are projected to increase over time--which will influence the benefits paid by defined-contribution pension schemes--a reference year must be chosen. For the purpose of this study, 2040 is used, because it provides a sufficiently long contribution period over which to approx- imate steady-state conditions. The first critical task is to investigate levels of income replacement across a relevant spectrum of income. Income is represented as a percentage (50­200 percent) of average earnings. The second task is to investigate the impact on income replacement of differences in the duration, timing, and density of an individual's contribution history (density refers to the per- centage of time an individual actually contributes over a given period). To facilitate the presentation of these multidimensional results, we compute replacement rates as a function of the age at which an individual exits the labor market. They are presented separately for full-career and partial- career workers. Replacement rates for full-career workers. Full-career workers are exam- ined first.14 Gross replacement rates indicate that the earnings-related pension scheme is progressive (that is, it provides higher levels of income replacement to people with lower levels of preretirement income) (fig- ure 9.1). The level of gross income replacement provided to someone earning half the average wage is 17 percentage points higher than that provided to someone earning twice the average wage. The situation does not change significantly when taxes are taken into consideration. The impact of taxes on net replacement rates increases as incomes rise (figure 9.2). As a result, the proportional increase from gross to net replacement rates is slightly higher for high-income workers than for low-income workers. Examination of net replacement rates by income level suggests that pensions for full-career workers in Slovenia can be considered adequate (figure 9.3).15 Net replacement rates decrease with income, indicating the progressiveness of the scheme. Replacement rates for all levels of preretirement income are higher than the middle benchmark, and replacement rates for low- and middle-income workers are higher than 282 Adequacy of Retirement Income after Pension Reforms Figure 9.1 Sources of Gross Replacement Rates in Slovenia, by Income Level 80 70 60 (percent) 50 rate 40 30 replacement 20 10 gross 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage first pillar (defined benefit) Source: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. Figure 9.2 Sources of Net Replacement Rates in Slovenia, by Income Level 100 90 80 70 (percent) 60 rate 50 40 30 20 replacement 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage first pillar (defined benefit) taxes Source: APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. the highest benchmark. Given that benefits for the lowest full-career workers significantly exceed the poverty line, the objective of poverty alleviation is being met. Income replacement levels for all workers are higher than regional and world averages, suggesting that the Slovenian pension system provides relatively generous benefits. Slovenia 283 Figure 9.3 Net Replacement Rates for Male Full-Career Workers in Slovenia, Europe and Central Asia, and the World 100 high benchmark 90 80 70 (percent) 60 middle benchmark rate 50 40 low benchmark replacement 30 net 20 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage Slovenia average Europe and Central Asia average world average poverty line (percentage of average earnings) Source: Authors' calculations based on World Bank 2007 and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. Replacement rates for partial-career workers. Not everyone works from age 20 to the statutory retirement age. Many individuals enter and exit the labor force (often at different ages and for different periods of time) and earn different wages while working (figure 9.4). To examine the ade- quacy of benefits for partial-career workers, we consider three stylized cases.16 These cases include career type A (someone entering the labor force at age 25 who works continuously for a period of years before leaving the workforce at some point between the ages of 50 and 70 and then claims a benefit); career type B (identical to career type A, except that the worker enters the workforce at age 30 and leaves no earlier than age 55); and career type C (identical to career type A, except that the individual contributes in only three years out of four while in the labor force). In cases where the withdrawal from the formal labor market occurs before the statutory retirement age, the pension is claimed (and the replacement rate calculated) only at the later age. For withdrawals after the statutory retirement age, the ages coincide. 284 Adequacy of Retirement Income after Pension Reforms Figure 9.4 Net Replacement Rates for Male Middle-Income Partial-Career Workers in Slovenia, by Career Type and Exit Age high benchmark 100 world and Europe and Central Asia average 90 80 70 (percent) middle benchmark 60 rate 50 low benchmark 40 poverty line (percentage of average earnings) 30 replacement 20 net 10 0 45 50 55 60 65 70 exit age from labor market career type A: career type B: career type C: entry age--25, entry age--30, entry age--25, contribution contribution contribution density--100% density--100% density--75% Source: Authors' calculations based on World Bank 2007 and the APEX model. Note: Figure shows projected replacement rate for 2040 as approximation of steady-state conditions. See text for description of career types. Several conclusions can be drawn from figure 9.4. First, all workers receive levels of income replacement higher than the poverty line. Second, leaving the workforce early is costly, because partial-career workers who retire long before the retirement age may receive income replacement slightly below the middle of the three benchmarks. Third, entering the workforce later in life is costly, because entering the workforce at age 30 results in a net replacement rate 3­10 percentage points lower than entering at age 25. Fourth, working intermittently is costly, because the net replacement rate of a worker who contributes three out of four years will be 4­18 percentage points lower than that of a worker who con- tributes continuously. In all cases, net replacement rates grow faster the longer someone continues to work past normal retirement age for up to four years after which replacement rates become flat. The increase in replacement rates beyond the normal retirement age becomes steeper, Slovenia 285 because of higher accrual rates awarded for up to four years of delayed retirement. This implies that the scheme provides an incentive to defer retirement. Although career type A and B workers can attain the 60 per- cent benchmark before reaching the normal retirement age, career type C workers must work until the normal retirement age to attain this benchmark. To replace 80 percent of their preretirement earnings, career type A workers must work for two years past the normal retirement age. Career type B and C workers cannot attain the 80 percent bench- mark, even if they work until age 70. Fiscal Sustainability The sustainability of a pay-as-you-go first-pillar pension scheme is best evaluated in actuarial terms by estimating the scheme's actuarial deficit as the difference between its assets and liabilities. If a large actuarial deficit exists, the scheme is financially unsustainable and needs policy actions that increase its assets, reduce its liabilities, or both. A good proxy for the actuarial deficit is the difference between the present value of the scheme's expected future revenues (that is, contributions and other income) and expected future expenditures (that is, benefit payments, administrative costs, and other expenses) over an extended projection period. The difference between these two values represents an unfunded liability (sometimes referred to as a financing gap) on the public-sector balance sheet. Because this study is also concerned with the time path of revenues and expenditures (and the resulting balance across the projec- tion period ending in 2050), this more pragmatic approach has been taken. Projections of expenditures, revenues, and deficits are presented on the basis of available postreform fiscal projections. Reforms reduced the deficits of the first-pillar pension scheme from 4.5 percent of GDP in 1999 to 1.5 percent of GDP in 2005. The system remains unsustainable in the long term, however. Pension expenditures are projected to stabilize at about 11.2 percent of GDP for the next few years before rising gradually to 18.5 percent of GDP by 2050, driven by the wage indexation of pensions and the aging of the population.17 Revenues are projected to remain flat, at about 10.0 percent of GDP over the same period, resulting in a deficit of 8.8 percent by 2050 (figure 9.5). The aging of the Slovene population is driving these projections. Like other countries in the region, Slovenia will experience rapid aging over the next few decades as a result of low fertility rates and increased life expectancy. As a result, the old-age dependency ratio is projected to increase from 23.8 percent in 2005 to 60.3 percent by 2050 (figure 9.6). 286 Adequacy of Retirement Income after Pension Reforms Figure 9.5 Projected Fiscal Balance of Slovenia's Public Pension Scheme, 2005­50 20 15 10 GDP of 5 0 percentage ­5 ­10 ­15 2005 2010 2020 2030 2040 2050 expenditures revenues deficit Sources: European Commission 2007b; Institute of Macroeconomic Analysis and Development 2007. Figure 9.6 Projected Old-Age Dependency Ratio in Slovenia, 2005­50 70 60 (percent) 50 ratio 40 30 20 dependency 10 old-age 0 2005 2010 2020 2030 2040 2050 year Source: Reiterer 2008. What options exist for restoring the system to fiscal balance? Unfortunately, for policy makers, the options are limited. Revenues can be increased by raising the contribution rate. Alternatively--or in addi- tion, because the options are not mutually exclusive--expenditures can be reduced by cutting benefits, increasing the minimum number of years required to become eligible for benefits, or delaying the payment of benefits by raising the retirement age further. Because raising the contribution rate could threaten competitiveness and will likely strengthen incentives for tax evasion, it is typically not embraced by policy makers. This leaves policy makers with limited options: cutting Slovenia 287 benefits, tightening eligibility conditions, or raising the retirement age. An informal analysis suggests that retirement ages for men and women will have to be increased to at least 70 by 2050 to restore long-term fis- cal balance.18 Restoring sustainability may reduce the adequacy of ben- efits provided to future beneficiaries. If retirement ages are left unchanged and the current structure of the system is retained, further cuts in benefits--on the order of a 56 percent reduction in the average benefit provided under the first pillar--will be required for the system to become sustainable (figure 9.7). If benefits are adjusted to maintain a fiscal balance to the current level, full-career workers will receive replacement rates 27­34 percentage points lower in 2050 than they receive today, depending on their relative level of prere- tirement income. Two observations emerge from comparing these new (and lower) net replacement rates against the three benchmarks. First, a 56 percent reduc- tion in benefits would result in levels of income replacement for full-career workers that are below the poverty line, indicating that the pension scheme would not achieve its poverty alleviation objective for low-income Figure 9.7 Net Replacement Rates for Male Workers in Slovenia before and after Benefit Adjustment 100 90 80 high benchmark 70 (percent) 60 middle benchmark rate 50 40 low benchmark 30 replacement 20 net 10 0 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 earnings levels as multiple of average wage net replacement rate adjusted net replacement rate poverty line (percentage of average income) Source: Authors' calculations based on the APEX model. 288 Adequacy of Retirement Income after Pension Reforms workers. Second, the same reduction in benefits would likely frustrate the objective of smoothing consumption for all but high-income full-career workers. Replacement rates for middle-income workers are below the mid- dle 60 percent benchmark, while rates for low-income workers are sub- stantially below the 80 percent benchmark. This last observation is subject to three caveats. First, this analysis applies only to full-career workers, while the average worker now contributes for only 27­30 years, substantially less than the 43 years defining a full career. Contributing to the pension scheme for 33 years (10 years short of a full career) reduces net income replacement by 6 percentage points for low- income workers and 12 percentage points for high-income workers. As a result, reducing benefits to restore the scheme to fiscal balance would make it very difficult for the scheme to effectively smooth consumption for virtually all partial-career workers and for most full-career low- and middle-income workers. Second, if benefit cuts are combined with fur- ther increases in the retirement age, the cuts need not be as large to restore fiscal balance. Third, all workers have the option of saving outside of the first-pillar pension scheme. To increase income replacement by 1 percentage point, for example, a full-career worker would need to save only about 0.5 percent of his or her earnings from age 40 to the current age of retirement.19 Conclusions The reforms of 2000 improved the fiscal condition of Slovenia's pension system, but they failed to make the first-pillar scheme sustainable. Levels of net replacement for full-career workers remain well above 80 percent-- quite high relative to regional and global benchmarks for workers of all income levels (see chapter 1).As in other countries, however, workers with less than full careers--because they left the workforce before reaching the retirement age, worked intermittently, or have gaps in their employment history--risk receiving a level of income replacement closer to the lower benchmark of 40 percent. The generosity of replacement rates at retire- ment is enhanced by generous indexation practices, which fully link pen- sions to wages. Moreover, the reforms failed to adequately address the aging of the population, which will steadily raise the ratio of pensioners to workers over time. As a result, the pension scheme is projected to gener- ate a deficit equivalent to 8.8 percent of GDP by 2050. To restore the long-term fiscal balance of the pension scheme without recourse to general revenue financing, policy makers could raise retirement Slovenia 289 ages further--rough estimates suggest an age well above 70 by 2050. To realize the full fiscal impact of this measure, income replacement must be maintained at levels now associated with the current retirement age. Increasing the retirement age in step with increases in life expectancy at retirement is a natural choice, both for individuals and for policy makers, but it requires cross-sectoral policy reforms to enable elderly workers to continue to participate in the labor market.20 Other options for restoring fiscal balance include cutting benefits at retirement or reducing the generosity of benefit indexation (or some combination thereof). Rough estimates suggest that average first-pillar replacement rates at retirement would have to fall some 30 percentage points. Moving from wage indexation to price indexation would reduce average first-pillar benefits by some 25 percentage points. Reducing ben- efits by such magnitudes would, however, likely compromise the objective of smoothing lifetime consumption for full-career low- and middle-income workers when measured against the three benchmarks used here. It would also likely frustrate the objective of alleviating poverty among eld- erly people with low preretirement income, because levels of income replacement would fall below the poverty line. This suggests that policy makers in Slovenia will have to either sup- port the pension scheme with funds from the general budget or enact further reforms to restore fiscal balance. Given the magnitude of pro- jected deficits, restoring fiscal balance through benefit cuts alone may frustrate the objective of smoothing consumption. Raising retirement ages further and cutting benefits may be the only viable option. Even if benefits are reduced, individuals have the option of supplementing their benefits by saving outside the mandatory first-pillar scheme. Given that participation in the voluntary third-pillar scheme remains small, an opportunity exists for broadening the reach of the scheme. Notes 1. The number of pensioners rose 18 percent between 1990 and 1992, as a result of the provision in the pre-1992 law that allowed older contributors to pay lower rates for missing years of contributions and then claim old-age pensions (Novak 2004). To ease the economic transition, the government shifted some of the costs to the pension system by allowing mass early retire- ment (Verbic, Majcen, and Nieuwkoop 2005). 2. There are two mandatory closed-end second-pillar schemes: one for employees working in hazardous occupations and (since April 2004) one for government 290 Adequacy of Retirement Income after Pension Reforms employees. Both are financed by small contributions paid by employers only. The scheme for government employees was introduced to address wage demands by public employees while still complying with the Maastricht criteria in the run-up to adopting the euro. Because these two schemes do not apply to all employees, they are excluded in the system characteriza- tion and analysis. 3. Mandatory defined-contribution schemes provide early retirement benefits for certain occupations, including miners, police officers, military personnel, and artists, who are assumed to work under harsh work conditions. Upon reaching age 58, qualified workers are entitled to an old-age pension from the pay-as- you-go scheme in addition to a pension from their occupational scheme. Employers are required to pay additional contributions of 4.2­12.6 percent (depending on the occupation) to finance these occupational pensions. The schemes are administered by the state-owned pension management fund. 4. In 2006, for example, the revalorization coefficient applied to wages earned in 1990 was only 78 percent of actual nominal wage growth over the period 1990­2005. 5. The minimum pension assessment base is set nominally. It was about 64 percent of the average wage in 2003. 6. For a critical view of the regulatory framework, see Berk (2008). 7. This provision was reportedly used largely by men in recent years. 8. The accrual rate is 3 percent the first year, 2.6 percent the second year, 2.2 percent the third year, and 1.8 percent the fourth year after the retire- ment age of 63 for men and 61 for women. If an individual does not fulfill the entire qualifying period and continues to work after the retirement age, the accrual rate is 0.3 percent per month the first year, 0.2 percent the second year, and 0.1 percent the third year. 9. The reduction is 0.3 percent per month at age 58, 0.25 percent per month at age 59, 0.2 percent per month at age 60, 0.15 percent per month at age 61, and 0.1 percent per month at age 62. 10. In Chile, for instance, 70 percent of retirees from the mandatory public pension system own their home, which is a form of savings (see Valdés-Prieto 2008). 11. These benchmarks approximate the standards developed by the International Labour Organization (ILO) and the Council of Europe (1990). ILO Convention 102 of 1952 sets a minimum benefit equal to 40 percent of the reference wage for married men of pensionable age. This amount was subse- quently increased to 45 percent in 1968. The European Code of Security of 1990 sets a minimum standard for members of the Council of Europe equal to 65 percent for married people of a specific age. 12. As a proxy for the poverty line, this study uses 35 percent of the average net wage, which very broadly approximates a US$2.25-a-day poverty line Slovenia 291 converted into national currency, adjusted for purchasing power parity, expressed relative to the national average net wage, and averaged across the eight study countries. Such an approach enables valid comparisons to be made across the sample (see chapter 1 of this volume). 13. The APEX model was developed by Axia Economics, with funding from the OECD and the World Bank. The model codes detailed eligibility and benefit rules for first- and second-pillar schemes based on available public information that has been verified by country contacts. Because the details of the rules sometimes change on short notice (and limited public disclosure), the calculations presented here should be considered as best approximations only. 14. For this analysis, a full career is defined as continuous employment from age 20 to the current normal retirement age of 63. 15. Replacements rates are simulated for an unmarried male working a hypothet- ical career path under the assumption that real wage growth is 2 percent, inflation is 2.5 percent, the rate of return on invested assets is 3.5 percent, and the worker retires at the statutory retirement age. 16. Only male middle-income, partial-career workers are examined. Although replacement rates across income levels vary, the results of the analysis are similar. 17. For projections of different policy scenarios using an overlapping genera- tional model, see Majcen and Verbic (2008). 18. This estimate is based on the World Bank's baseline demographic projec- tions. It assumes that everyone over age 70 receives a pension, everyone age 20­70 contributes, and all pensioners receive the replacement rate awarded to the median worker (62 percent). 19. This estimate is based on the assumption that real wage growth is 2 percent, the net real rate of return on invested assets is 3.5 percent, and benefits (from both the unfunded and the funded pillars) are price indexed. 20. See Holzmann, MacKellar, and Repansek (2009) for a conference volume that addresses these issues for the countries of southeastern Europe. Bibliography Albreht, T., M. Cesen, D. Hindle, E. Jakubowski, B. Kramberger, V. Kerstin Petric, M. Premik, and M. Toth. 2002. Healthcare Systems in Transition: Slovenia. World Health Organization Regional Office for Europe, Copenhagen. Berk, A. 2008. "Contemporary Issues and Challenges in a Supplementary Pension System: The Case of Slovenia." In Pension Reform in Southeastern Europe: Linking to Labor and Financial Market Reforms, ed. R. Holzmann, L. MacKellar, and J. Repansek, 209­25. Washington, DC: World Bank. 292 Adequacy of Retirement Income after Pension Reforms Council of Europe. 1990. European Code of Social Security (Revised). Rome. European Commission. 2007a. Mutual Information System and Social Protection (MISSOC) database. http://ec.europa.eu/employment_social/. ------. 2007b. "Pension Schemes and Projection Models in EU-25 Member Countries." European Economy Occasional Paper 37, Economic Policy Committee and Directorate General for Economic and Financial Affairs, Brussels. GVG (Gesellschaft für Versicherungswissenschaft und -gestaltung) and European Commission. 2003. Study on the Social Protection Systems in the 13 Applicant Countries: Slovenia Country Report. Cologne. Holzmann, R., and R. Hinz. 2005. Old-Age Income Support in the 21st Century. Washington, DC: World Bank. Holzmann, R., L. MacKellar, and J. Repansek, eds. 2009. Pension Reform in South- eastern Europe: Linking to Labor and Financial Market Reforms. Washington, DC: World Bank. Institute of Macroeconomic Analysis and Development. 2007. Pension Expenditures and Revenues. Government of Slovenia, Slovenia Country Fische, Economic Policy Committee Aging Group, Ljubljana. Insurance Supervision Agency. 2006. Annual Report. Ljubljana. ILO (International Labour Organization). 1952. ILO Convention 102. Geneva: ILO. ------. 1967. ILO Convention 128. Geneva: ILO. ------. 2004. The Collection of Pension Contributions: Trends, Issues, and Problems in Central and Eastern Europe. Subregional Office for Central and Eastern Europe, Budapest. Majcen, B., and M. Verbic.. 2009. "Slovenian Pension System in the Context of Upcoming Demographic Development." In Pension Reform in Southeastern Europe: Linking to Labor and Financial Market Reforms, ed. R. Holzmann, L. MacKellar, and J. Repansek, 73­88. Washington, DC: World Bank. Novak, A. 2004. Pension Reform in Slovenia. University of Glasgow. OECD (Organisation for Economic Co-operation and Development). 2001. Ageing and Income: Financial Resources and Retirement in Nine OECD Countries, Paris: OECD. Reiterer, A. 2008. Population Development and Age Structure in Southeastern Europe until 2050. World Bank, Washington, DC. U.S. Social Security Administration. 2006. Social Security Systems throughout the World: Europe. Washington, DC: Social Security Administration. Valdés-Prieto, S. 2008. Designs for the First-Pillar Pensions and the 2008 Chilean Reform. http://editorialexpress.com/cgi-bin/conference/download.cgi?db_ name=SECHI2008&paper_id=130. Slovenia 293 Verbic, M., B. Majcen, and R. van Nieuwkoop. 2005. Sustainability of the Slovenian Pension System: An Analysis with an Overlapping-Generations General Equilibrium Model. Ljubljana: Institute of Economic Research. Whitehouse, E. 1999. Tax Treatment of Funded Pensions. Washington, DC: World Bank. World Bank. 2007. Pensions Panorama. Washington, DC: World Bank. World Bank and IIASA (International Institute for Applied Systems Analysis). 2001. "The Slovenian Pension System and Reform." Paper presented at the World Bank­IIASA conference, Vienna. WHO (World Health Organization). 2008. Database. http://www.who.int/ research/en/. Index Boxes, figures, notes, and tables are indicated by b, f, n, and t following the page number. A Albania, 7 Analysis of Pension Entitlements across age for receiving benefits Countries (APEX) model, 39­40, See also headings starting with 57n9 "replacement rates" annuities Bulgaria, 19, 26, 49, 69, 72, 85 Bulgaria, 69 Croatia, 26, 49, 94, 96, 98, 100, 111 Croatia, 96, 99 Czech Republic, 26, 117, 120, 125, 136, Czech Republic, 123 137, 140­41 disability payments, 30 general discussion of, 5, 20­24t, 26, 34, Hungary, 151, 154, 158, 160 49, 51 Poland, 185, 191, 193 Hungary, 19, 26, 49, 154, 160, 168­69, Slovak Republic, 243, 250 171, 174, 176n6 Armenia, 7 Poland, 26, 49, 182, 193 assessment of performance of recommendation to increase, 55 pension system Romania, 26, 49, 212, 216, 220, Bulgaria, 73­84 234­35 Croatia, 100­110 Slovak Republic, 26, 49, 246, 250, 263 Czech Republic, 127­41 Slovenia, 19, 26, 49, 268, 270, 273, 275, general discussion of, 37­52 287, 289 Hungary, 162­73 Agency for the Supervision of Financial Poland, 195­204 Services (Croatia), 97­98 recommendations to improve age-related social pension. See pension performance, 56 income supplement; social Romania, 223­34 assistance benefits Slovak Republic, 252­62 Agricultural Social Insurance Fund Slovenia, 278­88 (Poland), 192 295 296 Index Averting the Old-Age Crisis (World Bank), 6 contribution ceilings in, 20t, 26 Azerbaijan, 7 contribution rates in, 20t, 25 cutting benefits to increase fiscal sustainability in, 51, 82­83 B deferment of retirement in, 83 benchmarks for adequacy analysis. deficits anticipated in, 62, 83, 85 See benefit adequacy defined-benefit scheme in, 61 benefit adequacy early retirement in, 63, 72 Bulgaria, 62, 73­81 farmers' pensions in, 70 Croatia, 102­9 guaranteed minimum income program Czech Republic, 129­38 in, 65, 67, 72 general discussion of, 37­48 indexation in, 27, 47f, 75, 85 Hungary, 163­69 life expectancy in, 77, 85 indexation and, 27 lump-sum payments in, 69 Poland, 195­203 motivation for reform in, 62­64 reforms not focusing on, 2 fiscal balance prior to reform, 4f, Romania, 223­30 61­62, 63f, 63t Slovak Republic, 252­59 occupational pension schemes in, 63, Slovenia, 278­84, 290n12 67­69, 86n4 benefit calculations, 20­24t, 26 old-age and system dependency ratios, benefit indexation. See indexation 62, 64f best years of wages as basis for benefits pension fund management companies Hungary, 157 in, 69 Poland, 191 point systems for pension calculation in, Slovak Republic, 246 67, 71 Slovenia, 273 police pensions in, 72 Bosnia, 7 poverty alleviation in, 53, 78, 83 Bulgaria, 61­85 self-employed in, 70 age for receiving benefits in, 26, 49, 69, social assistance benefits in, 17t, 53 72, 85 structure of benefits in, 71­73 annuities in, 69 disability benefits, 28t, 69, 72t, assessment of performance of pension 72­73, 86n8 system in, 73­84 old-age benefits, 20t, 71­72, 86n6, benefit adequacy, 62, 73­81 86n9 fiscal sustainability, 49, 50f, 62, survivor benefits, 30, 31t, 73, 74t, 86n8 81­84, 82f, 84f, 87n16 tax exemptions in, 69 replacement rates for full-career universal pension schemes in, 67­69, workers, 40, 41f, 43, 43f, 52, 86n4 78­79, 78­80f, 87n14 women and pensions in, 49, 67, 69, 71, 83 replacement rates for partial-career workers, 45f, 79­81, 81f C benefit calculations in, 20t, 26 characteristics of pension system in, 8t, capital gains treatment 64­73 Poland, 192 earnings-related schemes, 65, 67­69, Slovak Republic, 16 68t, 70­71 Central Administration of National Pension health care system, 36t, 65, 69­70, Insurance (Hungary), 159 86n3 Central Registry of Insured People noncontributory scheme, 67 (REGOS, Croatia), 97­98 pillar design, 12t, 65­70, 66t characteristics of pension system voluntary scheme, 35t, 65, 69, 70t, Bulgaria, 8t, 64­73 85, 86n4 Croatia, 8t, 91­100 Index 297 Czech Republic, 117­27 fiscal balance prior to reform, 4f, general discussion of, 10­37 90, 90t Hungary, 8t, 150­62 old-age dependency ratio in, 91, 91f Poland, 9t, 184­95 pension fund management companies Romania, 9t, 213­23 in, 91, 98 Slovak Republic, 9t, 242­52 point systems for pension calculation in, Slovenia, 269­77 7, 11, 89, 91, 94, 98­99, 110, Chile 112nn3­4 financial market development in, 6 poverty alleviation in, 53, 107 home ownership in, 56n6 replacement of defined-benefit scheme pillar model of, 7, 10 in, 11, 89, 91, 94, 110 contribution ceilings, 20­24t, 25­26, 38 social assistance benefits in, 53 contribution rates, 20­24t, 25, 51 state-provided matching contributions Council of Europe, minimum benefit set in, 111 by, 56n7 structure of benefits in, 96­100 Croatia, 89­111 disability benefits, 28t, 99t, 99­100 age for receiving benefits in, 26, 49, 94, old-age benefits, 20t, 98­99, 96, 98, 100, 111 112nn3­6 annuities in, 96, 99 survivor benefits, 31t, 100, 101t assessment of performance of pension tax deductions in, 34, 96 system in, 100­110 women and pensions in, 49, 94, 98, 100, benefit adequacy, 102­9 112n2 fiscal sustainability, 49, 50f, 109­10, Croatian Institute for Health Insurance 110f, 112nn16­17 (HZZO), 96­97 replacement rates for full-career Croatian Pension Insurance Institute, 97, 100 workers, 40, 41, 41f, 42, 43f, 52, Croatian Tax Administration, 97 105f, 105­7, 106f, 107f, cutting benefits to increase fiscal 112nn11­12 sustainability replacement rates for partial-career Bulgaria, 51, 82­83 workers, 45f, 107­9, 108f, 110­11, Czech Republic, 51, 140 112nn13­15 general discussion of, 51, 54, 55 benefit calculations in, 20t, 26 Hungary, 51, 171 characteristics of pension system in, 8t, Romania, 51, 231, 232, 235 91­100 Slovak Republic, 261, 262­63 earnings-related schemes, 94, 95t, Slovenia, 51, 286­87, 289 97­98 Czech National Bank, 123 health care system, 36t, 96­97 Czech Republic, 2­3, 115­43 noncontributory scheme, 94 age for receiving benefits in, 26, 117, pillar design, 12t, 92­97, 93t 120, 125, 136, 137, 140­41 voluntary scheme, 35t, 94­96, 96t annuities in, 123 contribution ceilings in, 20t, 25 assessment of performance of pension contribution rates in, 20t, 25 system in, 127­41 deficits anticipated in, 89, 109, 111 benefit adequacy, 129­38 early retirement in, 98, 108­9, 112n15 fiscal sustainability, 49, 50f, 138­41, flat-rate benefit formula in, 20t, 98, 139f, 144nn16­17 106­7 replacement rates for full-career guaranteed minimum income program workers, 40, 41, 41f, 42, 43, 43f, in, 17t 52, 132f, 132­33, 133f, 134f, 140, indexation in, 27, 42, 47f, 48, 102 141f, 142, 144nn11­12 life expectancy in, 92, 99, 112n2 replacement rates for partial-career lump-sum payments in, 96 workers, 45f, 133­38, 135f, 136f, motivation for reform in, 90­91 137f, 142 298 Index benefit calculations in, 21t, 26 Hungary, 171, 175 characteristics of pension system in, Poland, 193, 203 117­27 Romania, 231, 235 earnings-related schemes, 120­22, Slovak Republic, 246, 250, 261, 264 121t, 125, 143nn4­5 Slovenia, 276, 285 health care system, 36t, 124­25 deficits anticipated noncontributory scheme, 120, 143n2 See also fiscal sustainability pillar design, 7, 12t, 118­25, 119t Bulgaria, 62, 83, 85 voluntary scheme, 35t, 122­23, 123t Croatia, 89, 109, 111 contribution ceilings in, 21t, 25 Czech Republic, 116, 138, 142 contribution rates in, 21t, 25 Hungary, 147, 170, 171, 173, cutting benefits to increase fiscal 177nn16­17 sustainability in, 51, 140 Poland, 181, 182, 204 deferment of retirement in, 126, 135 Romania, 230­31, 234 deficits anticipated in, 116, 138, 142 Slovak Republic, 239, 240, 260­61, 263 early retirement in, 117, 120, 126 Slovenia, 267, 268, 288 flat-rate benefit formula in, 21t, 116, defined-benefit schemes 117, 120, 143n3 See also pillar design guaranteed minimum income program Bulgaria, 61 in, 17t, 118, 120 Croatia, replacement of, 11, 89, 91, 94, indexation in, 27, 47, 47f, 129 110 life expectancy in, 139 Czech Republic, 11, 116, 120 lump-sum payments in, 123, 143n6 general discussion of, 11 motivation for reform in, 116­17 Hungary, 11, 147, 148, 154, 173 fiscal balance prior to reform, 4f, 117, Poland, replacement of, 11, 188, 205 117t Romania, replacement of, 11, 211, 213, occupational pension schemes in, 117 216, 234 old-age and system dependency ratios Slovak Republic, replacement of, 11, 241 in, 116, 139f, 139­40, 144nn14­15 Slovenia, 11, 270 parametric reforms in, 115, 117, 141­42 defined-contribution schemes pension fund management companies See also fully-funded defined-contribution in, 123 scheme; notional defined- poverty alleviation in, 53, 133, 138, 140 contribution (NDC) scheme prepaid expenditure taxes in, 16 benefit calculations in, 26 self-employed in, 124, 125 general discussion of, 13 social assistance benefits in, 53 vesting in, 25 state-provided matching contributions delaying retirement. See deferment of in, 122­23, 143, 143n7 retirement structure of benefits in, 125­27 demographics. See old-age and system disability benefits, 28t, 126t, 126­27 dependency ratios old-age benefits, 21t, 125­26, 143n2 dependents. See survivor benefits survivor benefits, 31t, 127, 128t disability benefits tax deductions in, 122­23 Bulgaria, 28t, 69, 72t, 72­73, 86n8 tax exemptions in, 122 Croatia, 28t, 99t, 99­100 weakness of parametric reform in, 142 Czech Republic, 28t, 117, 126t, 126­27 women and pensions in, 120, 125, 140 general discussion of, 27, 28­29t, 30 Hungary, 28t, 160­61, 161t Poland, 29t, 194t, 194­95 D Romania, 29t, 212, 218­19, 220­21, 221t deferment of retirement Slovak Republic, 29t, 248, 250­51, 251t Bulgaria, 83 Slovenia, 29t, 276t, 276­77 Czech Republic, 126, 135 vesting and, 27, 28­29t Index 299 E Hungary, 4f, 148, 149t, 150f Poland, 4f, 182, 183f, 183t early retirement Romania, 4f, 212, 213t Bulgaria, 63, 72 Slovak Republic, 4f, 240, 241f, 241t Croatia, 98, 108­9, 112n15 Slovenia, 4f, 268, 268t Czech Republic, 117, 120, 126 fiscal sustainability Hungary, 148, 160, 168 Bulgaria, 49, 50f, 62, 81­84, 82f, 84f, Poland, 193, 206n5 87n16 Romania, 212, 213, 220, 234 Croatia, 49, 50f, 109­10, 110f, Slovak Republic, 246, 250 112nn16­17 Slovenia, 268­69, 289n1, 290n3 Czech Republic, 49, 50f, 138­41, 139f, earnings-related schemes 144nn16­17 Bulgaria, 65, 67­69, 68t, 70­71 general discussion of, 48­52, 50f Croatia, 94, 95t, 97­98 Hungary, 49, 50f, 169­73, 170f, Czech Republic, 120­22, 121t, 125, 177nn18­19 143nn4­5 indexation and, 27 general discussion of, 13, 38 as motivation for reform in, 1­2, 3 Hungary, 154­57, 155­56t, 159­62 Poland, 49, 50f, 203­4, 204f Poland, 188­91, 189­90t, 192­93, Romania, 49, 50f, 230­33, 231f, 206n5 236nn13­15 Romania, 216­17, 217t, 219­20, 236n6 Slovak Republic, 49, 50f, 259­62, 260f, Slovak Republic, 245­46, 247t, 248­49, 265n10 264nn2­4 Slovenia, 49, 50f, 285­88, 286f, Slovenia, 272t, 272­73, 275, 290nn4­5 291nn17­19 taxation and, 13 flat-rate benefit formula Estonia, 8t Croatia, 20t, 98, 106­7 European System of National Accounts, 54 Czech Republic, 21t, 116, 117, 120, exempt-exempt-taxed (EET) regime, 15b 127, 143n3 See also tax exemptions French system as example for point system, 7 F full-career workers, replacement rates for. farmers' pensions See replacement rates for Bulgaria, 70 full-career workers general discussion, 6 fully-funded defined-contribution scheme Hungary, 159 Hungary, 148, 149, 154, 173 Poland, 192 Poland, 185, 205 Romania, 219 Romania, 211, 213, 218, 234 Slovak Republic, 248 Slovak Republic, 239, 246 Slovenia, 274 Slovenia, 267 financial instruments available for individ- ual investment, 6, 39 G Financial Supervision Commission (Bulgaria), 71, 86n2 General Health Insurance Fund Financial Supervision Commission (Czech Republic), 124 (Poland), 191, 193 German system as example for point financial system reform, 6, 10 system, 7 financing gap, defined, 48 guaranteed minimum income program fiscal balance prior to reform Bulgaria, 65, 67, 72 Bulgaria, 4f, 61­62, 63f, 63t Croatia, 17t Croatia, 4f, 90, 90t Czech Republic, 17t, 118, 120 Czech Republic, 4f, 117, 117t general discussion of, 16 general discussion of, 3, 4f Poland, 17t, 191, 193 300 Index Romania, 18t, 214, 216 defined-benefit scheme in, 11, 147, 148, Slovak Republic, 18t, 245 154, 173 early retirement in, 148, 160, 168 H farmers' pensions in, 159 fully-funded defined-contribution hazardous occupations, workers in scheme in, 148, 149, 154, 173 Bulgaria, pension schemes for, 63 indexation in, 27, 157, 163, 174­75, Slovenia, 289n2 176n8 health care system life expectancy in, 175n2, 176n6 Bulgaria, 36t, 65, 69­70, 86n3 "lost generation" in, 174 Croatia, 36t, 96­97 lump-sum payments in, 158, 160 Czech Republic, 36t, 124­25 motivation for reform in, 148­50 general discussion of, 13, 34, 36t, 37, fiscal balance prior to reform, 4f, 148, 56n5 149t, 150f Hungary, 36t, 158­59 old-age and system dependency ratios Poland, 36t, 192, 206n6 in, 149, 150f Romania, 36t, 219 parametric reforms in, 148, 149, 154, 173 Slovak Republic, 36t, 248 pension fund management companies Slovenia, 36t, 274­75 in, 159 Health Insurance Fund (Hungary), 159 poverty alleviation in, 53, 168, 172 Health Insurance Institute of Slovenia, 274 prepaid expenditure taxes in, 16 Herzegovina, 7 self-employed in, 158, 159 home ownership in Chile, 56n6 social assistance benefits in, 53 Hungarian Financial Supervisory Authority, structure of benefits in, 159­62 158, 159 disability benefits, 28t, 160­61, 161t Hungary, 147­75 old-age benefits, 17t, 21t, 159­60 age for receiving benefits in, 49 survivor benefits, 32t, 161­62, 162t annuities in, 151, 154, 158, 160 tax deductions in, 158 assessment of performance of pension tax exemptions in, 154 system in, 162­73 unemployment in, 175 benefit adequacy, 163­69 women and pensions in, 49, 154, 160, fiscal sustainability, 49, 50f, 169­73, 171, 175n2 170f, 177nn18­19 replacement rates for full-career I workers, 41, 41f, 42, 43, 43f, 53, 166­67f, 166­68, 172f increasing age of retirement. See deferment replacement rates for partial-career of retirement workers, 45f, 168­69, 169f indexation benefit calculations in, 21t, 26 Bulgaria, 27, 47f, 75, 85 characteristics of pension system in, 8t, Croatia, 27, 42, 47f, 48, 102 150­62 Czech Republic, 27, 47, 47f, 129 earnings-related schemes, 154­57, general discussion of, 3, 20­24t, 26­27, 155­56t, 159­62 46­48, 47f health care system, 36t, 158­59 Hungary, 27, 157, 163, 174­75, 176n8 noncontributory scheme, 154 Poland, 27, 47, 47f, 48, 197, 207n7 pillar design, 7, 12t, 151­59, 152­53t recommendation of, 54­55 voluntary scheme, 35t, 157­58, 158t Romania, 46, 47f, 57n12, 224, 234, 235 contribution ceilings in, 21t, 26 Slovak Republic, 27, 47f, 48, 246, 253, cutting benefits to increase fiscal 263 sustainability in, 51, 171 Slovenia, 27, 47, 47f, 48, 273, 278­79, deferment of retirement in, 171, 175 289 deficits anticipated in, 147, 170, 171, individual retirement accounts, 39, 52 173, 177nn16­17 See also voluntary scheme Index 301 Insurance Act (Slovenia), 274 military pensions Insurance and Pension Funds Supervisory Poland, 192 Commission (Poland), 191 Romania, 219 Insurance Supervision Agency (Slovenia), Slovenia, 290n3 274 minimum-income guarantee program. international investment of Kosovo See guaranteed minimum income scheme, 10 program International Labour Organization (ILO), Ministry of Finance (Czech Republic), 123, minimum benefit set by, 7, 56n7 124, 144n13 Ministry of Labor, Family and Social Affairs K (Slovenia), 275 Ministry of Labor and Social Affairs Kazakhstan, 7, 8t, 10 (Czech Republic), 125, 144n13 Kosovo, 8t, 10 Ministry of Social Affairs (Czech Kyrgyz Republic, 7 Republic), 124 Montenegro, 7 L motivation for reform last years of wages as basis for benefits See also fiscal balance prior to reform See also best years of wages as basis for Bulgaria, 62­64 benefits Croatia, 90­91 Romania, 212, 216 Czech Republic, 116­17 Latin American examples, 6, 7 general discussion of, 3­10 See also Chile Hungary, 148­50 Latvia, 7, 8t Poland, 182­84 life annuities. See annuities Romania, 212­13, 235n1 life expectancy at retirement Slovak Republic, 240­42 Bulgaria, 77, 85 Slovenia, 268­69, 289n1 Croatia, 92, 99, 112n2 Czech Republic, 139 N general discussion of, 40, 41­42 National Bank of Slovakia, 249 Hungary, 175n2, 176n6 National Health Fund (Poland), 192 Poland, 205 National Health Insurance Fund (Bulgaria), Romania, 234, 235n5 70 Slovak Republic, 246, 261 National Health Insurance Fund Slovenia, 285 (Romania), 219 Lithuania, 8t National Health Insurance Fund "lost generation," 55, 174 Administration (NHIFA, Hungary), lump-sum payments 158 Bulgaria, 69 National House of Pensions (Romania), Croatia, 96 220 Czech Republic, 123, 143n6 National Revenue Agency (Bulgaria), 71 disability benefits, 30 National Social Security Institute (NSSI, Hungary, 158, 160 Bulgaria), 70­71, 86n7 Poland, 191, 195 noncontributory scheme Romania, 221, 223 Bulgaria, 67 Slovak Republic, 246, 250 Croatia, 94 survivor benefits, 30 Czech Republic, 120, 143n2 voluntary old-age benefits, 34 Hungary, 154 Poland, 188 M Romania, 216 Macedonia, 9t Slovak Republic, 245 mandatory funded schemes. See pillar design Slovenia, 270­72 302 Index notional defined-contribution (NDC) Croatia, 2 scheme Czech Republic, 115, 117, 141­42 in Poland, 2, 11, 42, 57n13, 181­82, general discussion of, 2, 7 185, 188, 191, 193, 205, 206n2 Hungary, 148, 149, 154, 173 in Sweden, 7 Romania, 2 Slovak Republic, 2, 260, 263 O Slovenia, 267, 269 partial-career workers, replacement rates occupational pension schemes for. See replacement rates for Bulgaria, 63, 67­69, 86n4 partial-career workers Czech Republic, 117 pay-as-you-go structure. See pillar design Poland, 206n3 Pension and Old-Age Round Table, 177n16 Romania, 212, 219 pension fund management companies Slovenia, 290n3 Bulgaria, 69 old-age allowance in Hungary, 17t, 154 Croatia, 91, 98 old-age and system dependency ratios Czech Republic, 123 Bulgaria, 62, 64f Hungary, 159 Croatia, 91, 91f Poland, 193 Czech Republic, 116, 139f, 139­40, Romania, 219 144nn14­15 Slovak Republic, 249 general description of problem related Slovenia, 274, 290n3 to, 2, 5, 5f pension income supplement in Slovenia, Hungary, 149, 150f, 175nn3­5 53, 270 Poland, 182­83, 184f Pensions and Disability Insurance Institute Romania, 213, 231­32, 232f (PDII, Slovenia), 275 Slovak Republic, 240­41, 242f performance assessment. See assessment of Slovenia, 285, 286f, 290nn8­9 performance of pension system old-age benefits pillar design Bulgaria, 20t, 71­72, 86n6, 86n9 Bulgaria, 12t, 65­70, 66t Croatia, 20t, 98­99, 112nn3­6 Croatia, 12t, 92­97, 93t Czech Republic, 21t, 125­26, 143n2 Czech Republic, 7, 12t, 118­25, 119t general discussion of, 16, 17t, 19­27 general discussion of, 7, 10, 11­13, 12t Hungary, 21t, 159­60 Hungary, 7, 12t, 151­59, 152­53t Poland, 22t, 193­95, 207n7 Poland, 7, 12t, 185­92, 186­87t Romania, 22­23t, 220 Romania, 12t, 214­19, 215t Slovak Republic, 23t, 249­50 Slovak Republic, 12t, 243­48, 244t Slovenia, 24t, 275­76 Slovenia, 7, 12t, 270­75, 271t, 289n2 Orbán, G., 177n16 point systems for pension calculation Organisation for Economic Co-operation Bulgaria, 67, 71 and Development (OECD) Croatia, 7, 11, 89, 91, 94, 98­99, 110, countries 112nn3­4 health insurance programs in, 37 general discussion of, 2 study on living standards for people 10 Romania, 7, 11, 211, 213, 216, 220, years older and people 15 years 235nn3­4 younger than normal retirement Slovak Republic, 7, 11, 241­42, 250 age, 39 Poland, 181­206 orphans. See survivor benefits age for receiving benefits in, 26, 49, 182, 193 P annuities in, 185, 191, 193 Palotai, D., 177n16 assessment of performance of pension parametric reforms system in, 25, 195­204 adjustment in second half of 1990s, 3 benefit adequacy, 195­203 Index 303 fiscal sustainability, 49, 50f, 203­4, structure of benefits in, 193­95 204f disability benefits, 29t, 194t, 194­95 replacement rates for full-career old-age benefits, 22t, 193­95 workers, 40, 41f, 43, 43f, 53, old-age benefits in, 207n7 200­201, 200­202f survivor benefits, 32t, 195, 196t replacement rates for partial-career women and pensions in, 25, 49, 191, workers, 45f, 201­3, 203f, 205 193, 194, 205 benefit calculations in, 22t, 26 police pensions "bridging pensions" in, 206n5 Bulgaria, 72 characteristics of pension system in, 9t, Poland, 192 184­95 Slovak Republic, 249 earnings-related schemes, 188­91, Slovenia, 290n3 189­90t, 192­93, 206n5 poverty health care system, 36t, 192 alleviation noncontributory scheme, 188 Bulgaria, 53, 78, 83 pillar design, 7, 12t, 185­92, 186­87t Croatia, 53, 107 voluntary scheme, 35t, 191­92, 192t Czech Republic, 53, 133, 138, 140 contribution ceilings in, 22t, 26 general discussion of, 16, 53 contribution rates in, 22t, 25 Hungary, 53, 168, 172 deferment of retirement in, 193, 203 Poland, 53, 198, 200­201 early retirement in, 193, 206n5 Romania, 53, 227, 232, 235 farmers' pensions in, 192 Slovak Republic, 53, 257, 259, 261 fully-funded defined-contribution Slovenia, 53, 282, 287 scheme in, 185, 205 poverty line, proxy for, 57n8 guaranteed minimum income program promoting formal sector employment, 54 in, 17t, 191, 193 purpose of reforms, 1­2 indexation in, 27, 47, 47f, 48, 197, 207n7 R life expectancy in, 205 lump-sum payments in, 191, 195 ratio of population age 65 and older military pensions in, 192 to population age 15­64. motivation for reform in, 182­84 See old-age and system dependency fiscal balance prior to reform, 4f, 182, ratios 183f, 183t replacement rates for full-career workers notional defined-contribution (NDC) Bulgaria, 40, 41f, 43, 43f, 52, 78­79, scheme in, 2, 11, 42, 57n13, 78­80f, 87n14 181­82, 185, 188, 191, 193, 205, Croatia, 40, 41, 41f, 42, 43f, 52, 105f, 206n2 105­7, 106f, 107f, 112nn11­12 occupations and pension schemes in, Czech Republic, 40, 41, 41f, 42, 43, 43f, 206n3 52, 132f, 132­33, 133f, 134f, 140, old-age and system dependency ratios 141f, 142, 144nn11­12 in, 182­83, 184f general discussion of, 40­43, 41f, 43f, pension fund management companies 52­53, 57n10 in, 193 Hungary, 41, 41f, 42, 43, 43f, 53, police pensions in, 192 166­67f, 166­68, 167f, 172f, 173, poverty alleviation in, 53, 198, 200­201 176nn12­15 replacement of defined-benefit scheme Poland, 40, 41f, 43, 43f, 53, 200­201, in, 11, 188, 205 200­202f, 207nn11­13 "Security through Diversity" reform Romania, 40, 41f, 43, 43f, 53, 227, program, 205, 206n1 227­28f, 234, 236n11 self-employed in, 192, 206n6 Slovak Republic, 40, 41f, 42, 43, 43f, 53, social assistance benefits in, 53, 188 255­58, 256­57f, 261, 262f, 265n8 304 Index Slovenia, 41, 41f, 43, 43f, 53, 273, indexation in, 46, 47f, 57n12, 224, 234, 281­83, 282­83f, 287f, 235 291nn14­15 last years of wages as basis for benefits replacement rates for partial-career workers in, 212, 216 Bulgaria, 45f, 79­81, 81f life expectancy in, 234, 235n5 Croatia, 45f, 107­9, 108f, 110­11, lump-sum payments in, 221, 223 112nn13­15 military pensions in, 219 Czech Republic, 45f, 133­38, 135f, 136f, motivation for reform in, 212­13, 235n1 137f, 142 fiscal balance prior to reform, 4f, 212, general discussion of, 43­46, 45f, 53, 213t 57n11 old-age and system dependency ratios Hungary, 45f, 168­69, 169f, 173­74 in, 213, 231­32, 232f Poland, 45f, 201­3, 203f, 205 pension fund management companies Romania, 45f, 229f, 229­30, 236n12 in, 219 Slovak Republic, 45f, 258f, 258­59, point systems for pension calculation in, 265n9 7, 11, 211, 213, 216, 220, Slovenia, 44, 45f, 283­85, 284f 235nn3­4 Romania, 211­35 poverty alleviation in, 53, 227, 232, 235 age for receiving benefits in, 26, 49, 212, replacement of defined-benefit scheme 216, 220, 234­35 in, 11, 211, 213, 216, 234 assessment of performance of pension self-employed in, 219 system in, 223­34 structure of benefits in, 220­23 benefit adequacy, 223­30 disability benefits, 29t, 212, 218­19, fiscal sustainability, 49, 50f, 216­17, 220­21, 221t 217t, 219­20, 236n6 old-age benefits, 22­23t, 220 replacement rates for full-career survivor benefits, 32t, 221­23, 222t workers, 40, 41f, 43, 43f, 53, 227, tax deductions in, 218 227­28f, 234, 236n11 women and pensions in, 49, 216, 218, replacement rates for partial-career 220, 234 workers, 45f, 229f, 229­30, 236n12 Romanian Private Pension System benefit calculations in, 22­23t, 26 Supervision Commission, 220 characteristics of pension system in, 9t, Russia, 7, 9t 213­23 earnings-related schemes, 216­17, S 217t, 219­20, 236n6 health care system, 36t, 219 Securities Market Agency (Slovenia), 274 noncontributory scheme, 216 "Security through Diversity" reform pillar design, 12t, 214­19, 215t program (Poland), 205, 206n1 voluntary scheme, 35t, 218t, 218­19 self-employed persons and pensions contribution ceilings in, 22­23t, 25 Bulgaria, 70 contribution rates in, 22­23t, 25 Czech Republic, 124, 125 cutting benefits to increase fiscal general discussion of, 6 sustainability in, 51, 231, Hungary, 158, 159 232, 235 Poland, 192, 206n6 deferment of retirement in, 231, 235 Romania, 219 deficits anticipated in, 230­31, 234 Slovak Republic, 248­49 early retirement in, 212, 213, 220, 234 Slovenia, 274, 275 farmers' pensions in, 219 Serbia, 7 fully-funded defined-contribution Slovak Republic, 239­64 scheme in, 211, 213, 218, 234 age for receiving benefits in, 26, 49, 246, guaranteed minimum income program 250, 263 in, 18t, 214, 216 annuities in, 243, 250 Index 305 assessment of performance of pension self-employed in, 248­49 system in, 252­62 social assistance benefits in, 53 benefit adequacy, 252­59 structure of benefits in, 249­52 fiscal sustainability, 49, 50f, 259­62, disability benefits, 29t, 248, 250­51, 260f 251t replacement rates for full-career old-age benefits, 23t, 249­50 workers, 40, 41f, 42, 43, 43f, 53, survivor benefits, 33t, 251t, 251­52 255­58, 256­57f, 262f tax deductions in, 246 replacement rates for partial-career tax exemptions in, 16, 245 workers, 45f, 258f, 258­59 women and pensions in, 49, 246, 249 benefit calculations in, 23t, 26 Slovenia, 3, 267­89 best years of wages as basis for benefits age for receiving benefits in, 26, 49, 268, in, 246 270, 273, 275, 287, 289 capital gains treatment in, 16 assessment of performance of pension characteristics of pension system in, 9t, system in, 278­88 242­52 benefit adequacy, 278­84, 290n12 earnings-related schemes, 245­46, fiscal sustainability, 49, 50f, 285­88, 247t, 248­49 286f, 291nn17­19 health care system, 36t, 248 replacement rates for full-career noncontributory scheme, 245 workers, 41, 41f, 43, 43f, 53, pillar design, 12t, 243­48, 244t 281­83, 282­83f, 287f, voluntary scheme, 35t, 246­48, 248t 291nn14­15 contribution ceilings in, 23t, 26 replacement rates for partial-career contribution rates in, 23t, 25 workers, 44, 45f, 283­85, 284f, cutting benefits to increase fiscal 291n16 sustainability in, 261, 262­63 benefit calculations in, 24t, 26 deferment of retirement in, 246, 250, best years of wages as basis for benefits 261, 264 in, 273 deficits anticipated in, 239, 240, characteristics of pension system in, 260­61, 263 269­77 early retirement in, 246, 250 earnings-related schemes, 272t, farmers' pensions in, 248 272­73, 275 fully-funded defined-contribution health care system, 36t, 274­75 scheme in, 239, 246 noncontributory scheme, 270­72 guaranteed minimum income program pillar design, 7, 12t, 270­75, 271t, in, 18t, 245 289n2 indexation in, 27, 47f, 48, 246, 253, 263 voluntary scheme, 35t, 273­74, 274t life expectancy in, 246, 261 contribution ceilings in, 24t, 25 lump-sum payments in, 246, 250 contribution rates in, 24t, 25 motivation for reform in, 240­42 cutting benefits to increase fiscal fiscal balance prior to reform, 4f, 240, sustainability in, 51, 286­87, 289 241f, 241t deferment of retirement in, 276, 285, old-age and system dependency ratios, 289 240­41, 242f deficits anticipated in, 267, 268, 288 pension fund management companies defined-benefit scheme in, 11, 270 in, 249 early retirement in, 268­69, 289n1, point systems for pension calculation in, 290n3 7, 11, 241­42, 250 farmers' pensions in, 274 police pensions in, 249 fully-funded defined-contribution poverty alleviation in, 53, 257, 259, 261 scheme in, 267 replacement of defined-benefit scheme hazardous occupations workers in, in, 11, 241 289n2 306 Index indexation in, 27, 47, 47f, 48, 273, structure of benefits 278­79, 289 See also disability benefits; health care life expectancy in, 285 system; old-age benefits military pensions in, 290n3 Bulgaria, 71­73 motivation for reform in, 268­69, Croatia, 96­100 289n1 Czech Republic, 125­27 fiscal balance prior to reform, 4f, 268, Hungary, 159­62 268t Poland, 193­95 occupational pension schemes in, Romania, 220­23 290n3 Slovak Republic, 249­52 old-age and system dependency ratios Slovenia, 275­77 in, 285, 286f, 290nn8­9 survivor benefits parametric reforms in, 267, 269 Bulgaria, 30, 31t, 73, 74t, 86n8 pension fund management companies Croatia, 31t, 100, 101t in, 274, 290n3 Czech Republic, 30, 31t, 127, 128t pension income supplement in, 53, 270 general discussion of, 30, 31­33t police pensions in, 290n3 Hungary, 30, 32t, 161­62, 162t poverty alleviation in, 53, 282, 287 Poland, 30, 32t, 195, 196t self-employed in, 274, 275 Romania, 32t, 221­23, 222t state pension in, 18t Slovak Republic, 33t, 251t, 251­52 structure of benefits in, 275­77 Slovenia, 33t, 277, 277t disability benefits, 29t, 276t, Sweden as example of NDCs, 7 276­77 Swiss indexation. See indexation old-age benefits, 24t, 275­76, 290nn8­9 T survivor benefits, 33t, 277, 277t taxation of contributions and benefits, women and pensions in, 49, 272­73, 13­16, 14t, 15b 275­76 See also tax deductions; tax exemptions social assistance benefits tax deductions See also pension income supplement in Croatia, 34, 96 Slovenia Czech Republic, 122­23 Bulgaria, 53, 67, 70 general discussion of, 34 Croatia, 53, 94 Hungary, 158 Czech Republic, 53, 120, 124 Romania, 218 general discussion of, 2, 16, 19 Slovak Republic, 246 Hungary, 53, 154 tax exemptions Poland, 53, 188 Bulgaria, 16, 69 Slovak Republic, 53, 245, 259 Czech Republic, 122 Social Insurance Agency (Slovak Republic), exempt-exempt-taxed (EET) regime, 249 15b Social Insurance Institution (ZUS, Poland), Hungary, 154 192­93 noncontributory scheme benefits, 56n2 social pensions in Bulgaria, 17t, 19 Slovak Republic, 16, 245 See also social assistance benefits taxed-exempt-exempt (TEE) regime, Social Security Administration (Czech 15b Republic), 125 Turkmenistan, 7 state pension in Slovenia, 18t, 270, 290n3 state-provided matching contributions U Croatia, 111 Czech Republic, 122­23, 143, 143n7 Ukraine, 7, 9t State Tax Collection Agency (Hungary), universal pension schemes in Bulgaria, 159 67­69, 86n4 Index 307 V Croatia, 49, 94, 98, 100, 112n2 Czech Republic, 120, 125, 140 vesting periods Hungary, 49, 154, 160, 171, disability benefits, 27, 28­29t 175n2 old-age benefits, 19, 20­24t, 25 Poland, 25, 49, 191, 193, voluntary scheme 194, 205 Bulgaria, 35t, 65, 69, 70t, 85, 86n4 Romania, 49, 216, 218, 220, 234 Croatia, 35t, 94­96, 96t Slovak Republic, 49, 246, 249 Czech Republic, 35t, 122­23, 123t Slovenia, 49, 272­73, 275­76 general discussion of, 1, 13, 34, 35t, 39, survivor benefits, 30 52, 53, 57n15 World Bank Hungary, 35t, 157­58, 158t, 175 assessment principles for evaluating Poland, 35t, 191­92, 192t pension systems, 37 Romania, 35t, 218t, 218­19 See also assessment of performance of Slovak Republic, 35t, 246­48, 248t pension system Slovenia, 35t, 273­74, 274t Averting the Old-Age Crisis, 6 conceptual framework of pension W systems by, 10 widows/widowers. See survivor benefits See also pillar design women and pensions influence on pension reform, Bulgaria, 49, 67, 69, 71, 83 56n1 ECO-AUDIT Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. · 9 trees The Office of the Publisher has chosen to · 6 million BTUs of total print Adequacy of Retirement Income after energy Pension Reforms in Central, Eastern, and · 6,069 lbs. of CO2 Southern Europe on recycled paper with 30 equivalent of green- percent post-consumer waste, in accordance house gases with the recommended standards for paper · 3,150 gallons of usage set by the Green Press Initiative, a non- waste water profit program supporting publishers in using · 405 pounds of fiber that is not sourced from endangered solid waste forests. For more information, visit www.greenpressinitiative.org. All countries in the former transition economies of Central, Eastern, and Southern Europe have undertaken public pension reforms of varying depth and orientation, often with the support of the World Bank. Although the reformed public pension schemes provide broad benefit adequacy, in most cases additional measures are needed to achieve fiscal sustainability in an aging society. Adequacy of Retirement Income after Pension Reforms in Central, Eastern, and Southern Europe: Eight Country Studies assesses the benefit adequacy of the reformed pension systems for eight countries--Bulgaria, the Czech Republic, Croatia, Hungary, Poland, Romania, the Slovak Republic, and Slovenia--to identify policy gaps and options. The authors identify the motivations for reform against the backdrop of the trend toward multi-pillar arrangements, document key provisions, and compare them in the context of the World Bank's five-pillar paradigm for pension reform. They then evaluate the sustain- ability and adequacy of reformed pension systems and provide recommendations to address gaps and take advantage of opportunities for further reforms. The case studies and summary suggest the following broad policy conclusions: · Fiscal sustainability has improved in most study countries, but few are fully prepared for the inevitability of population aging. · The linkage between contributions and benefits has been strengthened, and pension system designs are better suited to market conditions. · Levels of income replacement are generally adequate for all but some categories of workers (including those with intermittent formal sector employment or low lifetime wages), and addressing their needs requires initiatives that go beyond pension policy. · Further reforms should focus on extending labor force participation by the elderly to avoid benefit cuts that could undermine adequacy and very high contribution rates that could discourage formal sector employment. · More decisive financial market reforms are needed for funded provisions to deliver on the expectations of participants and keep funded pensions safe. This book will be of interest to policy makers, researchers, and everyone interested in the topic of pensions in the region, and beyond. ISBN 978-0-8213-7781-9 SKU 17781