78979 Povert y Reduction and Economic Management Unit | Europe and Central Asia Region Mitigating the Economic Impact of an Aging Population Options for Bulgaria September 2013 Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Poverty Reduction and Economic Management Europe and Central Asia Region Document of the World Bank CONTENTS Abbreviations and Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . vii Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ix Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxvii CHAPTER I. Demographic Change, Labor Supply and the Economic Life-Cycle. . 1 Demographic trends and projections . . . . . . . . . . . . . . . . . . . . 1 Labor force projections and scenarios . . . . . . . . . . . . . . . . . . . . 4 The Economic Life-Cycle . . . . . . . . . . . . . . . . . . . . . . . . . 6 Policy Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 CHAPTER II. Labor Market Policies to Mitigate the Impact of Population Aging. 15 Labor Force Participation by Population Groups . . . . . . . . . . . . . . 16 Policy Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 CHAPTER III. The Role of Education Policies . . . . . . . . . . . . . . . . 23 Bulgaria’s Basic Education in the Context of Demographic Change. . . . . . 24 Tertiary Education and Demographic Change . . . . . . . . . . . . . . . 26 Recent Reforms in the Education Sector . . . . . . . . . . . . . . . . . . 29 Improving Educational Outcomes among the Roma . . . . . . . . . . . . . 32 Policy Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 CHAPTER IV. Health Sector Reform in the Context of an Aging Population . . 37 Current Public Health Expenditure. . . . . . . . . . . . . . . . . . . . 38 Current Health-Sector Performance in the Context of Aging. . . . . . . . . 39 A Service-Delivery Agenda with a Focus on the Hospital Sector. . . . . . . 41 Policy Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 CHAPTER V. Long-Term Care Expenditures and Policy Options. . . . . . . . 49 Current Public Expenditures on Long-Term Care . . . . . . . . . . . . . . 50 Providing Long-Term Care Services . . . . . . . . . . . . . . . . . . . . 52 Financing of LTC Services. . . . . . . . . . . . . . . . . . . . . . . . 53 Policy Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 iii CHAPTER VI. The Effect of Demographic Change on the Pension System . . . . 57 Key Features of Bulgaria’s Pension System. . . . . . . . . . . . . . . . . 58 Impact of Recent Reforms. . . . . . . . . . . . . . . . . . . . . . . . 63 Projections of Fiscal Balance under the Baseline . . . . . . . . . . . . . . 66 Alternative Financing Mechanisms for a more Equitable Pension System. . . . 68 Reform Scenarios. . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Policy Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 CHAPTER VII. Macro-Fiscal Implications of Bulgaria’s Demographic Change. . 75 Key Baseline Assumptions . . . . . . . . . . . . . . . . . . . . . . . . 76 Key Findings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Alternative Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Annex I: Population Projections for Bulgaria. . . . . . . . . . . . . . . . . . 99 FIGURES Figure 1: One century of demographic change. . . . . . . . . . . . . . . . . . xii Figure 2: Working-age population and old-age dependency ratio. . . . . . . . xii Figure 3: Labor force participation rates and labor supply scenario by age-groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Figure 4: Bulgaria’s population at risk of poverty or social exclusion by age group. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xiv Figure 5: Bulgaria’s hospital expenditure – age profile. . . . . . . . . . . . . . xv Figure 6: Monthly distribution of number of Bulgarian hospital patients. . . . . xv Figure 7: Projected Baseline Fiscal Payg Balance (in percent of GDP). . . . . . . xvi Figure 8: Projections of GDP growth and debt-to-GDP ratios under different scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii Figure 9: Mandatory minimum length of paid maternity leave. . . . . . . . . xviii Figure 10: NEET rate among the youth (aged 15–24) in 2012. . . . . . . . . . . xix Figure 11: Average long-term care spending 2003–2007 (in % of GDP). . . . . . xx Figure I-1: One century of demographic change in Bulgaria. . . . . . . . . . . . 2 Figure I-2: Determinants of population growth. . . . . . . . . . . . . . . . . . . 3 Figure I-3: Change in age structure and dependency ratios. . . . . . . . . . . . . 4 Figure I-4: Labor force participation rates under different scenarios. . . . . . . . . 5 Figure I-5: Population and labor force projections under different scenarios . . . . . 6 Figure I-6: Bulgaria’s hospital expenditure-age profile . . . . . . . . . . . . . . . 7 Figure I-7: Bulgaria’s expenditure composition in comparison . . . . . . . . . . . 7 Figure I-8: Income and private consumption for selected countries. . . . . . . . . 9 Figure I-9: Gross savings and sources of household income. . . . . . . . . . . . . 10 Figure I-10: Growth and convergence . . . . . . . . . . . . . . . . . . . . . . . . 11 Figure II-1: Employment rates for 20–64 year-olds in 2010. . . . . . . . . . . . . 16 Figure II-2: NEET rate among 15–24 year-olds in 2009. . . . . . . . . . . . . . 17 Figure II-3: Employment rates among 50–64 year-olds. . . . . . . . . . . . . . . 17 Figure II-4: Discrimination perceptions . . . . . . . . . . . . . . . . . . . . . . . 18 Figure II-5: Labor Market Rates for Bulgarian Women and Men in 2009 (in percent). 18 iv | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure II-6: Employment rates for Roma and non-Roma neighbors . . . . . . . . 20 Figure II-7: Gap in tax wedge between high and low wage earners in 2008 . . . . 21 Figure III-1: Public education expenditures in percent of GDP . . . . . . . . . . . . 24 Figure III-2: Evolution of student numbers and ratios . . . . . . . . . . . . . . . . 25 Figure III-3: Enrolment rates of Roma in Bulgaria by age in 2011 . . . . . . . . . . 33 Figure IV-1: Public health expenditure growth and share in public expenditure. . . 39 Figure IV-2: Acute care hospital beds per 100thds . . . . . . . . . . . . . . . . . . 43 Figure IV-3: Typology of hospitals in Bulgaria. . . . . . . . . . . . . . . . . . . . 43 Figure IV-4: Monthly distribution of number of Bulgarian hospital patients. . . . . 44 Figure VI-1: Dependency rates and population pyramid. . . . . . . . . . . . . . . 58 Figure VI-2: Contributors and their age structure in 2011 . . . . . . . . . . . . . . 60 Figure VI-3: Pension contribution rates and spending in percent of GDP in 2010 . . 61 Figure VI-4: Financial performance of social insurance . . . . . . . . . . . . . . . . 62 Figure VI-5: Gross replacement rates. . . . . . . . . . . . . . . . . . . . . . . . 63 Figure VI-6: Retirement duration in 2010 . . . . . . . . . . . . . . . . . . . . . . 65 Figure VI-7: Disability and old age pensioners . . . . . . . . . . . . . . . . . . . . 65 Figure VI-8: Projected Baseline Fiscal Payg Balance (in percent of GDP). . . . . . . . 66 Figure VI-9: Projected replacement rate for an average old age pensioner. . . . . . 68 Figure VI-10: Scenario I: PAYG fiscal balance and female replacement rate . . . . . 70 Figure VI-11: Scenario II: Projected PAYG fiscal balance. . . . . . . . . . . . . . . . 71 Figure VI-12: Scenario III: Projected PAYG balance . . . . . . . . . . . . . . . . . . 71 Figure VII-1: Bulgaria’s capital flows in percent of GDP. . . . . . . . . . . . . . . . 77 Figure VII-2: Fiscal Projections under the Baseline Assumptions. . . . . . . . . . . . 81 Figure VII-3: Projections under TFP scenarios. . . . . . . . . . . . . . . . . . . . . 83 Figure VII-5: Labor force by education level under different scenarios . . . . . . . . 86 Figure VII-6: Projections under Education Scenarios. . . . . . . . . . . . . . . . . . 87 TABLES Table I-1: Decomposition of public spending on secondary education. . . . . . . . 8 Table III-1: Implications of demographic challenges for Bulgaria’s Education system. . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Table III-2: Growth of enrollments in tertiary education in Bulgaria . . . . . . . . 26 Table III-3: Global education responses to demographic shifts . . . . . . . . . . . 30 Table III-4: Europe 2020 education goals . . . . . . . . . . . . . . . . . . . . . . 31 Table III-5: Highest level of education attained, by ethnic group in percent . . . . 32 Table III-6: Recommendations for areas of strategic focus and collaboration . . . . 34 Table IV-1: Health Expenditure - Bulgaria and Comparator Countries (2010). . . . . 38 Table IV-2: Health Status Indicators of Bulgaria and comparable countries. . . . . . 40 Table IV-3: Use of preventive services in Bulgaria and selected EU countries. . . . 41 Table IV-4: Inpatient hospital discharges Bulgaria and select EU countries. . . . . 42 Table IV-5: Activity amenable to outpatient care in 5 multi-profile hospitals. . . . 46 Table V-1: Share of highly dependent people by age group in 2008 . . . . . . . . 50 Table V-2: Total LTC expenditures as a share of GDP in selected countries. . . . . 51 Table V-3: Government expenditure on LTC according to government Expenditure data (BGN). . . . . . . . . . . . . . . . . . . . . . . . 52 Table VI-1: Structure of the Bulgarian pension system . . . . . . . . . . . . . . . . 59 Table VI-2: Pension benefit types provided by Pillar 0 and I in 2011. . . . . . . . . 59 CONTENTS | v Table VI-3: Summary of main legislated reforms. . . . . . . . . . . . . . . . . . 64 Table VI-4: Projected pensioners as a percent of population above the retirement age . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Table VI-5: Old age pension levels in 2011. . . . . . . . . . . . . . . . . . . . . 67 Table VI-6: Countries with universal and basic pensions. . . . . . . . . . . . . . . 69 Table VI-7: Scenario III: Macroeconomic assumptions. . . . . . . . . . . . . . . . 71 Table VII-1: Baseline Assumptions. . . . . . . . . . . . . . . . . . . . . . . . . . 76 Table VII-2: Decomposition of Debt Dynamics under the Baseline Scenario. . . . . . 82 Table VII-3: Growth rates and public debt ratio under baseline and scenarios. . . . 89 Table VII-4: Comparison of Debt Dynamics across the Scenarios . . . . . . . . . . . 89 BOXES Box III1: Denmark’s strategy for lifelong learning. . . . . . . . . . . . . . . . . . . 56 Box VII1: Underlying models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 vi | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Acronyms and Abbreviations BGN Bulgarian Lev LTC Long-term Care BURS Bulgarian University Ranking LTCI Long-term Care Insurance System MEYS Ministry of Education, Youth CCPs Clinical Care Pathways and Social Sciences CEE Central and Eastern Europe MLSP Ministry of Labor and Social DRGs Diagnostic Related Groups Policy ECA Europe and Central Asia MoH Ministry of Health EU EU MoF Ministry of Finance EU10 Bulgaria, Czech Republic, NCDs Non-communicable Diseases Estonia, Hungary, Latvia, NEEA National Evaluation and Lithuania, Poland, Romania, Accreditation Agency Slovak Republic, Slovenia NEET Not in Employment, Education EU15 EU of 15 member states: or Training EU-12 plus Austria, Finland, NGO Non-Governmental and Sweden Organization EU25 EU of 25 member states: NHA National Health Accounts EU15 plus Czech Republic, NHIF National Health Insurance Fund Cyprus, Estonia, Hungary, NICE National Institute for Health Latvia, Lithuania, Malta, and Clinical Excellence, UK Poland, Slovakia, and Slovenia NMS New Member States EU27 EU of 27 member states: plus NSLL National Strategy of Bulgaria and Romania Life-Long Learning GDP Gross Domestic Product NTA National Transfer Account GNI Gross National Income NYS National Youth Strategy HEI Higher Education Institute OOP Out-of-Pocket Expenditure HFA Health For All PDL Positive Drug List HiT Health Systems in Transition PHE Public Health Expenditure ILO International Labour PIRLS Progress in International Organization Literacy Study LE Life Expectancy PISA Program for International LFP Labor Force Participation Students Assessments LLL Life-long Learning PROST Pension Reforms Options LOD Labor Offices Directorate Simulation Toolkit vii R&D Research and Development TVET Technical and Vocational QOF Quality and Outcomes Education Training Framework UN United Nations SDR Standard Death Rate USAID United States Agency for SHA System of Health Accounts International Development TFR Total Fertility Rate WHO World Health Organization TFP Total Factor Productivity Currency Equivalents Exchange Rate as of August 27, 2013 Currency Unit Bulgarian Lev US$ 1.00 BGN 1.47 Government Fiscal Year January 1–December 31 Vice President: Philippe Le Houerou Country Director: Mamta Murthi Sector Director: Yvonne Tsikata Sector Manager: Satu Kahkonen Task Team Leaders: Doerte Doemeland Zeljko Bogetic viii | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Acknowledgements This report responds to  a  request from the Vienna) and a note on Bulgaria’s private sav- Bulgarian Minster of  Finance from Octo- ings prepared by Samuel Munzele Maimbo ber 2012. It was prepared under the overall (Lead Financial Specialist, ECSF2). Olga leadership and guidance of  Satu Kahkonen Kupets (Ukraine National University “Ky- (Sector Manager, ECSP2). The team also iv-Mohyla Academcy”, Kiev) provided ex- benefited greatly from the support and guid- cellent research assistance. Chapter II has ance of  Mamta Murthi (Country Director, been prepared by Mohamed Ihsan Ajwad ECCU5), Peter Harrold (Former Country (Senior Economist, ECSH4) with inputs Director, ECCU5), Yvonne Tsikata Sector from Plamen Danchev (Education Special- Director, ECSPE) and Markus Repnik ist, ECSH2), Joost de Laat (Senior Econo- (Country Manager, ECCBG). The team mist Human Development, ECSH4) and would like to  thank staff at  the Ministry Carolin Geginat (Senior Private Sector De- of  Finance (MoF), in  particular Marinela velopment, FGIDB). Chapter III has been Petrova (Director of  Economic and Finan- written by Roberta Malee Bassett (Senior cial Policies), as  well as  the counterparts Education Specialist, ECSH2) with inputs from the Ministry of  Economy, Ministry from Plamen Danchev and Igor Kheyfets of  Labor and Social Policy, Ministry (Economist, ECSH2). Chapter IV has been of Health, the National Social Security In- composed by Agnès Couffinhal (Senior stitute and the National Statistical Institute Economist Health, ECSH1 with consider- for their excellent collaboration and invalu- able support from Técnicas de Salud, S.A. able feedback throughout the preparation (Spain) and Antoniya Dimova (Varna Uni- of  this report, largely provided during two versity of Medicine). The support of Petko workshops in  January  2013  in  April  2013, Salchev (National Center of Public Health hosted by the MoF. and Analysis), Dessislava Dimitrova and The task team was led by Doerte Do- Minchov Vichev (MoH) is gratefully ac- emeland (Senior Economist, DECOS) and knowledged. Chapter V has been compiled Zeljko Bogetic (Lead Economist, ECSP2). by Johannes Koettl with inputs from Georgi Chapter I has been written by Doerte Do- Shopov (Institute of Economics at Bulgarian emeland, Johannes Koettl (Senior Econo- Academy of Sciences). Chapter VI has been mist, ESCH4), Stella Ilieva (Senior drafted by Miglena Abels (Consultant, Economist, ECSP2) and Anil Onal (Consul- ECSH3) with inputs from Asta Zviniene tant, ECSP2) drawing on background pa- (Senior Social Protection Specialist, pers prepared by Anna Raggl (University of ECSH3). Chapter VII has been written by ix Harun Onder (Economist, PRMED) and Economist, ECSH3). The team would also Doerte Doemeland. The underlying model like to thank Roumeen Islam (Economic has been developed by Harun Onder with Adviser, ECSPE), Kaspar Richter (Lead contributions by Eduardo Ley (Lead Econo- Economist, ECSP3) and Gallina Vincelette mist, PRMED), Doerte Doemeland and (Senior Economist, ECSP2) for their impor- Pierre Pestieu (University of Liege), and data tant contribution at the concept note stage. and inputs from Miglena Abels, Agnès Stella Ilieva (Senior Economist, ECSP2) pro- Couffinhal, Plamen Danchev, Samir KC vided excellent feedback to the team during (University of Vienna) and Johannes Koettl. the drafting stage and was instrumental in The team benefited from helpful com- managing the coordination with the Bul- ments from the peer reviewers, Jorge Araujo garian Authorities. (Lead Economist, LCSPE), William Malo- The team also extends its gratitude ney (Lead Economist, DECMG) and Luigi to  Gabriela Calderon (Program Assistant, Giamboni (ECFIN, European Commis- DECOS), Andreina Clower (Program Assis- sion), and from comments and suggestions tant, ECSP2), Nancy Davies-Cole (Program from Eduardo Ley (Lead Economist, Assistant, ECSP2) and Adela Delcheva PRMED), Pedro Rodriguez (Lead Econo- Nachkova (Team Assistant, ECCBG) for ex- mist, ECSP2) and Anita Schwarz (Lead cellent administrative and logistics support. x | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Executive Summary Challenges the slow population growth before the 1990s as well as its steep decline over 1. Bulgaria is undergoing a profound the last two decades. Emigration alone socio-economic transformation contributed to a 10 percent decline of brought about by extraordinary the economically active population demographic change. Between 1950 since the 1990s. Bulgaria’s age structure and 1990, its population grew from 7.3 has changed radically (Figure 1). Its me- million to 8.8 million and then fell in dian age increased from 30.3 in 1960 to half the time to 7.5 million by 2010. 42.7 in 2012, the third-highest median Low birth rates, high mortality rates age in the EU, surpassed only by Ger- and significant emigration explained many and Italy. Figure 1: One century of demographic change 1950 2010 2050 100+ 95–99 90–94 85–89 80–84 75–79 70–74 65–69 60–64 Age Groups 55–59 50–54 45–49 40–44 35–39 30–34 25–29 20–24 15–19 10–14 5–9 0–4 400 300 200 100 100 200 300 400 400 300 200 100 100 200 300 400 400 300 200 100 100 200 300 400 0 0 0 Population in Thousands Female Male Source: Authors’ calculations based on data from the United Nations Population Division (2011). xi 2. Bulgaria is heading for the steepest As output per worker increases, fewer drop in the working-aging popula- workers will be effectively required to tion of any country, and it will im- pay for existing health sector, pension pose a heavy burden on the econo- and long-term care liabilities. Higher my. According to UN population growth is also likely to generate better projections, its labor supply is projected employment opportunities inducing to decline by up to 40 percent till 2050 more workers to participate in the labor (Figure 2). Its old-age dependency ratio, force and to stay in or move to Bulgaria. i.e. the share of elderly in the total pop- It is also required to raise household sav- ulation, is expected to double over the ings, which are needed to ensure that next four decades. By 2050, one in three the elderly can afford a decent standard Bulgarians is projected to be older than of living. 65 and only one in two Bulgarians will be of working age. Since the proportion 4. Bulgaria’s demographic change is of the population that works is a key de- however likely to depress produc- terminant of a country’s income level, tivity growth through three main its decline is likely to depress growth. channels: First, a low population densi- ty may reduce the exploitation of hu- 3. The higher productivity growth, man capital externalities and economies the easier it will be for Bulgaria to of scale. Second, aging may depress en- manage this demographic chal- trepreneurial activity and innovation. lenge. Given its declining working-age Third, there is a danger that population population, Bulgaria will have to rely aging will depress fiscal savings as age- on productivity growth to sustain sensitive expenditures such as health, growth in aggregate income. Productiv- pensions and long-term care will in- ity growth means that a country is able crease and tax revenue will decline, re- to produce more output with the same ducing the fiscal savings needed for pro- input factors, such as capital and labor. ductivity-enhancing public investments. Figure 2: Working-age population and old-age dependency ratio a. Working-Age Population Index b. Old-Age Dependency Ratio 110 55 100 50 90 45 40 80 35 70 30 60 25 50 20 2010 2015 2020 2025 2030 2035 2040 2045 2050 2010 2015 2020 2025 2030 2035 2040 2045 2050 Bulgaria EU15 EU10 (excl. Bulgaria) Source: Authors’ calculations based on United Nations Population Division (2011). Note: Index is calculated as 2010 equal 100. Working-age population is population aged 15–65. Old-age dependency ratio is calculated as the population age 65 as a share of total population. xii | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria 5. There may be some potential to education and health care. More flexible mitigate the impact of the declin- work arrangements would need to be ing working-age population by ex- promoted and work-places adapted to panding the work force. Women, el- the needs of an older work force.The ed- derly and young people have low Labor ucation sector would need to expand in- Force Participation (LFP) rates in Bul- vestment in lifelong learning to keep an garia. In 2012, the LFP rate of women aging work-force productive at a time of aged 15 to 24 was 25.3 percent. This fast technological change.The health sec- compares to an overall LFP rate of 30. 4 tor would need to become more efficient percent in Bulgaria, 33.1 in the EU10 to ensure that older people remain and 45.6 in the EU15 (Figure 3a). The healthy. Not only would the share of old- LFP rate of elderly Bulgarian women er workers increase but also the propor- aged 54 to 64 is nearly 10 percentage tion of workers from the Roma popula- points below the EU15 average. Still, tion. Roma already provide between 9 even under the most optimistic scenario and 19 percent of new labor market en- that assumes an increase of LFP rates of trants. This is likely to increase in the fu- women, elderly and young workers be- ture, requiring public services to adapt yond the highest LFP rates observed in better to their needs. Europe today, this decline cannot be fully stemmed. The most effective way 7. The high incidence of old-age pov- to stop the labor force from shrinking erty in Bulgaria and the currently further is to stanch emigration. rapid aging of its population also raise the question how an accept- 6. As Bulgaria ages, its work force will able standard of living for its elder- become radically different. The share ly can be funded. In 2011, nearly half of older workers is likely to increase. This of Bulgaria’s population of 7.3 million will require changes in the labor market, was poor or at risk of social exclusion, Figure 3: Labor force participation rates and labor supply scenario by age-groups a. Labor Force Participation Rates (2012) b. Labor Supply Under Alternative Scenarios 90 60,000 80 40,000 70 20,000 LFP Rates 60 0 50 –20,000 40 –40,000 30 –60,000 20 –80,000 15–24 24–54 54–64 Scenario I Scenario I Scenario III Scenario V (Constant LFP (Based on Eu (Convergence) (Maximum) Bulgaria (Female) Bulgaria (All) Population Projections) EU10 (excl. Bulgaria) EU15 15-39 40-64 65+ Source: Figure 3a: Authors’ calculations based on data from Eurostat (2012); Figure 3b: Authors’ simulations based on data from ILO (2012), United Nations Population Division (2011) and ILO (2012). Executive Summary | xiii Figure 4: Bulgaria’s population at risk of poverty or social exclusion by age group 80 In Percent of Population 60 40 20 0 2008 2009 2010 2011 Less than 16 16–24 25–49 50–64 65+ Total Source: Eurostat (2012). the highest percentage of all EU coun- tions, household savings are unlikely to tries. Poverty is much more prevalent increase significantly given continued among the elderly. Sixty percent of the high unemployment, weak growth and population aged 65 or above is at risk of a pervasive risk of poverty. poverty or social exclusion (Figure 4), posing challenges to fund an acceptable 8. Aging will affect the demand for standard of living for the elderly. Op- key public services. Demand for tions include relying on family or the health services and long-term care tend community; delaying retirement; in- to increase with age. Older people tend creasing public transfers to the old and to consume more health care as illness- encouraging savings. Relying on family es, chronic diseases and hospital visits is likely to become increasingly costly become more frequent in old age. Only for Bulgaria’s economy as the working- at a very old age does the health care de- age population shrinks. It will thus be- mand tend to decline (Figure 5). In Bul- come imperative to encourage savings, garia, people in their seventies have the including by further delaying retirement largest public health care costs per per- as life expectancy increases. In Bulgaria, son. Similarly, demand for long-term household savings have been negative care tends to increase steeply after peo- since 2005. Other European countries, ple turn 65 years of age. Only the old including those in Eastern Europe, have can benefit from old-age pensions. As achieved positive household savings the population ages, public expenditures rates; for example, Hungary and Poland on health, long-term care and pensions have maintained average rates of 10.6 can thus be expected to increase. On the and 7.3 percent, respectively, for the pe- other hand, demand for education tends riod 2005–10. Bulgaria has been imple- to be concentrated at younger ages. In menting some measures to encourage general, younger people are also more household savings in the long term. For likely to commit crimes. Population ag- example, it recently introduced a gradu- ing is therefore expected to reduce pub- al increase in the retirement age to reach lic expenditure for education as well as 65 for men and 63 for women by 2017 public order and safety. and 2020, respectively. The government has also put in place measures to increase 9. Population aging will place unprec- the employment rate. Despite these ac- edented demands on the provision xiv | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure 5: Bulgaria’s hospital expenditure – age profile Cost Index (Hospital Costs per 3.5 3.0 40–49 Year Old = 1) 2.5 2.0 1.5 1.0 0.5 0 0–9 10–19 20–29 30–39 40–49 50–59 60–69 70–79 80+ Source: Author’s calculations based on hospital data from MoH. of health care. The performance of tem abound. Bulgaria has a low cover- Bulgaria’s health system is lagging in age of preventive measures, a low use of many ways and Bulgarians are increas- outpatient services and high hospitaliza- ingly dissatisfied with it. This is likely to tion rates. Hospitals are very fragment- increase pressure to invest more in its ed, with a large number of hospitals various components. Though the health treating a low number of patients. sector has undergone a number of re- Among 275 Bulgarian hospital facilities forms over the past few years, these have ranked by number of patients discharged been unable to stem a rise in out-of per month, 193 discharge less than 13 pocket expenditures and improve the patients a day, which would be consid- health of Bulgarians relative to other ered the discharge of a small acute care EU countries. Inefficiencies in the sys- hospital in a country where service de- Figure 6: Monthly distribution of number of Bulgarian hospital patients 200 In a country where service delivery consolidation is advanced, small acute care hospital 150 might typically treat at least Number of Patients per Day 13 patients a day In Bulgaria, 103 hospitals discharge 100 less than 5 patients a day 50 In Bulgaria, 192 hospitals discharge less than 13 patients a day 0 #136 #123 #193 #17 #197 #244 #106 #34 #261 #84 #69 #47 #60 #159 #1 #99 #198 #92 #139 #238 #110 #144 #233 #124 #243 #177 #224 #90 #10 #95 #108 #44 #102 #153 #6 #256 #113 #112 #225 #201 #66 #59 #160 #105 #191 #158 #94 #93 #214 #152 #192 #275 #172 #223 #135 #42 #28 #179 #200 #241 #37 #240 #101 #36 #3 #264 #104 #140 #96 #12 #67 #236 #168 #119 #182 #181 #235 #46 #109 #63 #85 #170 #245 #41 #19 #255 #21 #9 #216 #97 #232 #267 #31 #204 #8 #263 #82 #132 #116 #213 #120 #269 #89 #30 #254 #111 #39 #76 #167 #57 #165 #24 #222 #11 #27 #217 #257 #156 #242 #231 #22 #162 #180 #50 #211 #239 #206 #14 #79 #71 #166 #13 #40 #122 #80 #52 #163 #16 Facilities Ranked By Number of Patients Source: Author’s calculations based on National Center of Health Informatics Data. Executive Summary | xv livery consolidation is advanced ising, some form of public risk-pool- (Figure 6). Hospital data also suggest ing—either in the form of tax-financed that around 20 percent of hospital ad- services or social insurance—will be missions correspond to conditions which necessary. Second, a sufficiently well- international standards suggest could be developed social sector for LTC services routinely treated on an outpatient basis. has important cost-efficiency effects on The aging population creates a pressing the health sector by shifting patients need to bring about improvements in with low-intensity care needs out of the the delivery of public health services. hospital sector. Third, given the expect- ed increase in the scarcity of labor, in- 10. Demand for long-term care (LTC) formal caregivers—in particular stay-at- services is bound to increase strong- home women—will be less and less ly with aging. As the Bulgarian popu- available, which will increase the de- lation ages, the number of dependent mand for formal care services. The pro- people in need of LTC will increase as vision of private and public long-term the number of potential caregivers de- care services, which is currently at a creases. This raises the question of who very low level in Bulgaria, would thus will take care of the elderly in the fu- need to be scaled up. ture. More and more responsibility is likely to fall upon the government. First, 11. Bulgaria’s pension system is pro- the costs of becoming dependent and jected to remain in deficit through- having to rely on the help of others can out the projection horizon. Though be very high and quickly impoverish el- other EU countries spend similar por- derly people. Since across the globe the tions of their GDP on pensions, have experience with private providers of similar retirement ages and comparable LTC insurance has not been very prom- benefits, Bulgaria’s low contribution Figure 7: Projected baseline fiscal PAYG balance (in percent of GDP) 0 –1 –2 Percentage –3 –4 –5 –6 –7 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040 2042 2044 2046 2048 2050 2052 2054 2056 2058 2060 2062 2064 2068 2070 2072 2074 PAYG Balance, %GDP Accounting for Government Contribution (Baseline) PAYG Balance, %GDP without Government Contribution (Baseline) Accounting for Additional Cost for Social Pensions (with Govt. Contribution) Accounting for Additional Cost for Social Pensions (without Govt. Contribution) Source: Author’s calculations using PROST. xvi | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria rates have led to declining revenues and efit from the government subsidies to the increasing subsidies from the state bud- pension fund, is projected to decline. get. In 2011, employee and employer Since citizens with pension rights are contributions financed only 53 percent likely to be wealthier, having worked in of pension expenditures. The remaining the formal sector for at least 37 years in liabilities were financed from general the case of women and 40 years in the revenue. Since 2009, the Government case of men, than those without pension has become a third insurer, paying con- rights, government subsidies to the pen- tribution equal to 12 percent of the total sion fund will benefit an increasingly contributory base. The Government al- small share of relatively well-off citizens. so covers the additional remaining defi- cit beyond the 12 percent contribution. 13. Simulations suggest that even under Going forward, the transfers from the optimistic assumptions, Bulgaria’s general government budget to the pen- demographic transformation will sion funds are projected to decline from exert steady fiscal pressures and de- 6.1 percent of GDP in 2012 to 4.9 per- press economic growth. As labor cent of GDP in the medium term before shrinks by close to 40 percent till 2050, climbing to 5.6 percent by 2050. real GDP growth is projected to slow down to 0.7 percent per year by the end 12. The current pension system also of the projection horizon under the base- raises equity concerns. Currently, the line scenario. Public expenditures as a Bulgarian pension system covers only share of GDP are expected to increase in about 55 percent of the working-age the long term as a result of spending for population. Coverage is expected to de- public health care, long-term care and cline as a result of high unemployment government transfers to the pension sys- rates in the medium-term and sporadic tem, resulting in an increase in Bulgaria’s contribution patterns due to a large in- debt-to-GDP ratio from 18 percent to formal work-force. As a result, the share 51 percent by the end of the projection of elderly with pension rights, who ben- horizon. Protracted primary deficits are Figure 8: Projections of GDP growth and debt-to-GDP ratios under different scenarios a. GDP per Capita Growth b. Debt-to-GDP Ratio in 2050 4.5 70 4.0 3.5 60 3.0 50 Percentage 2.5 Percentage 2.0 40 1.5 30 1.0 0.5 20 0.0 2012 2020 2025 2030 2035 2040 2045 2050 10 2015 0 Baseline Low TFP High TFP Low TFP Baseline EuroPOP High TFP High LFP High High LFP EuroPOP LFP High Education LFP Education Source: Authors’ simulations. Executive Summary | xvii the most prominent driver of the accu- geted at older people, youth and mulation of public debt over the projec- the Roma. Job-creation policies span a tion period, adding about 0.9 percentage wide range of demand- and supply-side points per year to the debt-to-GDP ratio. measures that must work in tandem. These policies include: (i) promoting 14. Scenario analyses show that the macroeconomic stability; (ii) improving government can play a significant the investment climate; (iii) refining la- role in mitigating these effects. Pol- bor market regulations; (iv) designing icies that induce higher labor force par- an employment-friendly tax system; and ticipation, promote productivity growth, (v) developing an adaptable workforce and provide better education outcomes through the education system. Still, could counteract significantly the fiscal since labor force participation rates in pressure arising from age-dependent fis- Bulgaria are particularly low for elderly cal expenditures. A low TFP scenario il- and youth, in particular elderly and lustrates that downward risks are signifi- young women, as well as minorities, cant. Maintaining strong productivity these general policies would need to be growth, which will entail active policies combined with policies specifically de- and planning, is key for maintaining signed to increase employment among debt at a sustainable level. these groups. Providing affordable alter- natives for child and elderly care servic- es could boost LFP rates of young and elderly women by reducing the oppor- Opportunities tunity cost of working outside the home. Since Bulgaria is among the countries 15. As Bulgaria’s labor force is project- with the longest paid maternity leave in ed to decline, measures to enhance the world (Figure 9), it could also con- productivity will become increas- sider reducing it while scaling up of ingly important to support labor child care services. Promoting flexible demand, competitiveness and work arrangements could draw more growth. Given the magnitude of Bul- youth, older workers and women into garia’s demographic challenge, improve- employment. Expanding part-time and ments in social sector policies alone are home-based employment opportunities unlikely to suffice. The implementation may encourage older workers to remain of productivity-enhancing policies will in the labor market until the stipulated be crucial for sustaining long-term retirement age. Early-childhood pro- growth. The latter could include policies grams for the Roma could help them in favor of strengthening Bulgaria’s busi- join the labor force as adults. ness climate, innovation and better access to infrastructure. These measures would 17. Since Bulgaria’s youth work and not only support productivity growth study less than those in other EU but also positively affect labor demand countries, additional measures tai- which combined with active policies in lored to them are important. More support of higher labor force participa- than one out of every five young Bulgar- tion will be necessary to stem the work- ians is classified as not in employment, force decline. education or training (NEETs). NEETs have usually dropped out of school with- 16. General labor policies would need out qualifications and are likely to be de- to be combined with policies tar- pendent on welfare programs. Reducing xviii | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure 9: Mandatory minimum length of paid maternity leave 450 400 350 300 250 200 150 100 50 0 United Kingdom Ireland Italy Finland Denmark Greece Austria France Netherlands Spain Belgium Germany Portugal Sweden Bulgaria Czech Republic Slovak Republic Hungary Estonia Poland Lithuania Romania Latvia Slovenia Brazil Korea Mexico Russian Switzerland Taiwan, China USA Japan EU15 Countries EU10 Countries Other Countries Source: World Bank (2012f). the number of NEETs could result in er youth as well as, back-to-school pro- significant productivity gains, fiscal sav- grams that would enable youth to ac- ings and poverty reduction in the medi- quire valuable skills. The experience um-term. In fact, it has been estimated from other countries suggests that the that the lack of labor market participa- impact of activation measures is mixed tion by NEETs costs Bulgaria about 2.3 and hence, there is a strong case for all percent of GDP. An effective youth measures to be evaluated. NEET strategy needs to focus on retain- ing young people in formal education 18. Reducing emigration, motivating and training, and be carefully targeted. Bulgarians to return from abroad Options include job placement services, and attracting immigrants will be training and remedial education for old- important to slow the workforce Figure 10: NEET rate among the youth (aged 15–24) in 2012 600 500 400 300 200 100 0 Austria Belgium Denmark Finland France Germany Ireland Italy Luxembourg Netherlands Portugal Spain Sweden United Kingdom Bulgaria Croatia Czech Republic Estonia Hungary Latvia Lithuania Poland Romania Slovakia Slovenia EU15 Average EU10 (excl. Bulgaria) Average Source: Eurostat (2012). Executive Summary | xix decline. Policies in support of employ- good example. The government could ment are also likely to reduce emigra- also explore reductions in security-re- tion and make Bulgaria more attractive lated public expenditures, which are for immigrants. Poorly functioning la- significantly above other EU countries. bor markets, weak public-service deliv- ery and an uncertain business environ- 21. Bulgaria has implemented signifi- ment have all been found to be significant cant education reforms over the drivers of emigration in Eastern Euro- past five years, paving the way for pean countries. High unemployment deeper changes. In recent years, Bul- rates, significant wage differentials be- garia has reduced the number of schools tween Bulgaria and neighboring coun- and teachers in line with its declining tries and the lack of favorable employ- student population. The quality of basic ment and business opportunities for education has improved and the avail- high-skilled people have all been found ability of tertiary education programs to be motives for Bulgarians to emi- has risen. However, Bulgaria needs to grate. Labor-market policies, education develop an effective program of life- policies and health-sector reforms, if long learning (LLL) and to reduce to- combined with policies in favor of im- day’s high drop-out rates, especially proving the business climate could add among the Roma and other disadvan- to Bulgaria’s labor supply by increasing taged groups. LLL opportunities are still LFP rates at the same time as reducing em- at an infant stage in Bulgaria and the igration. government could develop a national system of validation of LLL programs. It 19. To ensure that elderly people of could also create a comprehensive sys- have decent standard of living, en- tem involving all relevant stakeholders couraging household savings will to promote a good match between the be important in the medium-term. demands of employers and the availabil- The government could take steps to en- ity of LLL opportunities. To cut drop- courage savings by further increasing out rates, a first step would be the active the retirement age, encouraging supple- implementation of Bulgaria’s Strategy mentary voluntary pension insurance, for the Prevention and Reduction in the and scaling up financial information Share of Drop-Outs and Early Leavers and education. from the Education System (2013– 2020). Requiring all students to com- 20. Strengthening long-term fiscal plete general education program planning will be important to through age 16 would increase the avail- gauge the long-term policy impact ability of transversal skills. Continued of reforms in the context of a dras- efforts of improving the quality of and tically changing population. Institu- participation in education programs is tions need time to evolve, but Bulgaria‘s particularly important in an aging soci- population has aged very rapidly and ety as workers that have benefited from continues to do so. The government higher quality education are more likely thus needs to respond flexibly to chang- to be retrained in the course of their ing demand and prudently reduce pub- working lives. Bulgaria education is al- lic services as they become obsolete. ready well-supported by an extensive The introduction of per-capita spend- network of legal and strategic frame- ing in education in 2007, combined works to deal with in this new era of de- with the policy of protected schools, is a mographic decline. If the implementa- xx | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria tion of these recently developed policy physicians per capita is high, the number directives can focus on ensuring good of nurses per capita is by far the lowest in access to education for all and on im- the region. Solutions could include proving the quality of existing institu- planning for human resources, adapting tions, then the education sector is well- training, and addressing financial and poised to serve the entire population of other constraints in order to retain qual- Bulgaria well in the decades to come. ified staff in Bulgaria. 22. There are no easy or quick solu- 23. Providing financial protection to tions to prepare Bulgaria’s health all—so that illness does not lead to sector for population aging. Suc- poverty—is a core objective of the cessful reform requires strong vi- health-care system. Financial protec- sion and leadership. The fragmenta- tion provided by the National Health tion of the hospital sector is likely to lead Insurance Fund has significant gaps and to significant inefficiencies (Figure 11). out-of-pocket (OOP) expenditures are A key action would be the implementa- large in Bulgaria. Large OOP expendi- tion of a technically-driven hospital ra- tures on drugs, which are more prob- tionalization plan to support the reorga- lematic for the poor and the elderly, and nization of the hospital sector and the exclusions from the regular insurance creation of referral networks while system introduce inefficiencies. Patients maintaining physical access to care. In who forego care or do not manage their this context, the National Health Insur- existing conditions are at higher risk of ance Fund needs to be able to purchase complication and ultimately end up services selectively, i.e., it should be al- costing more to the public system. Thus, lowed to decide which entities it can improving financial protection is effi- contract. To support this process, infor- cient in addition to being equitable. mation on the quality of care needs to be generated, collected and publicized. 24. Bulgaria’s long-term care (LTC) is The implementation of a modern Diag- currently under-funded. Any public- nostic-Related Group (DRG) payment ly funded investments in building up system has been under consideration for more LTC and rehabilitative service fa- many years, but efforts to implement cilities, though, would need to be care- this have not succeeded. Accompanying fully evaluated. Because costs for LTC the introduction of DRGs with expen- can be catastrophically high, some form diture caps and the development of out- of risk-pooling is desirable. As pay-as- patient alternatives through concerted you-go financing mechanisms are un- efforts (regulatory and in terms of pay- sustainable in an aging society, earmark- ment systems) would avoid further esca- ing private savings of young workers for lation of costs. Strengthening alterna- LTC through, for example, financial tives to hospital-based care would LTC insurance is an option worth ex- require improving the capacity of pri- ploring. Finally, it is essential that Bul- mary-care professionals to manage the garia improves its data collection and re- prevailing burden of disease and to play cording practices on LTC services in an active role in the coordination of order to be able to develop effective and their patients’ care needs. The health- fiscally responsible policies in this area. care sector’s specialty mix better needs to adjust to better serve the population. 25. Though Bulgaria has implemented Although in Bulgaria the number of systemic and parametric reforms in Executive Summary | xxi Figure 11: Average long-term care spending 2003–2007 (in % of GDP) 4.0 3.5 3.0 2.5 2.0 1.5 0 Sweden Germany France Spain Portugal Slovenia Latvia Lithuania Poland Hungary Czech Republic Estonia Bulgaria Selected EU15 Countries Selected EU10 Countries Source: Eurostat (2010). Note: The 2005 value for Hungary (8.23 percent of GDP) was dismissed as an outlier. recent years, its pension system de- contribute during a shorter period of pends on large government subsi- time than men as they often stop work- dies. Increasing the contribution rate ing to raise children. Stipulating women levied on individual wages by 6 percent- to retire earlier than men results in age points would close the financing smaller benefits. In addition, an earlier gap. However, such a large increase may retirement age for women, who have a have potential negative effects on the la- longer life expectancy than men, further bor market, given the large size of Bul- increases the period of benefit receipt for garia’s informal economy. Bulgaria women relative to men. Facing inevita- could, however, consider a smaller in- bly smaller pensions in the future, Bul- crease in the PAYG contribution of garia should further develop its financial around 2 percentage points that would infrastructure, which would enable peo- reduce government subsidies without ple to save more for their retirement introducing major labor market distor- through different mandatory and volun- tions. Moreover, in the context of a rap- tary pension saving vehicles. id demographic contraction, pursuing both poverty alleviation and income re- 27. Strengthening disability certifica- placement inevitably leads to the com- tion processes would help to miti- pression of the income distribution gate a potential increase in disabili- among the elderly and reduces incen- ty claims due to high unemployment tives to contribute. and stricter eligibility conditions in the old age program. Disability 26. Several alternative financing mech- spending as a percentage of GDP more anisms could reduce state subsidies than doubled between 2001 and 2010. and improve redistribution. Equal- Despite the fact that the number of peo- izing retirement for men and women at ple who receive disability benefits has age 65 and increasing retirement ages been stabilized since 2006, it is likely that beyond 65, in the future, would be im- the pressure on the disability program portant to consider. Women tend to will intensify given further increases to xxii | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria the retirement age and the presently dif- al reform options can create syner- ficult labor market conditions. gies. Further increases in the retirement age, for example, would support higher 28. Avoiding further ad-hoc pension labor force participation among elderly increases would facilitate long-term workers, reduce public transfers to the fiscal planning. In the past, the Gov- pension fund and encourage household ernment has recurrently opted for in- savings. Well-targeted, strategic health- creasing pensions above the level pre- sector reforms could help improve citi- scribed by the law. Such practices make zens’ well-being, the efficiency of public it difficult to accurately project financ- health spending and increase the labor ing needs and ultimately jeopardize the supply, especially among elderly workers. long-term fiscal sustainability of the Well-designed investments in basic edu- pension fund. A pension reform is an in- cation could also raise labor force partici- herently difficult undertaking from a pation, make it easier to retrain workers at political standpoint and reducing pen- a later stage of their life, foster innovation sions, once they have been raised, is dif- and contribute to a healthier population. ficult to implement. Policymakers Reforms that exploit these synergies are would thus be well-advised in refrain- likely to be particularly effective. Still, ing from resorting to ad-hoc measures. one-off reforms are unlikely to be enough. A sustained reform commitment in all 29. While demographic change poses relevant areas will be required to mitigate different challenges and opportuni- the economic impact of Bulgaria’s demo- ties for each economic sector, sever- graphic change. Executive Summary | xxiii Summary of policy recommendations Challenge Urgent policy actions Medium-term recommendations Enhancing Growth Employment Policies Low employment rates Facilitate part-time employment by expanding Increase access to good quality and affordable among youth, older people, training opportunities, and increasing its use in childcare and elderly care. Roma and women the public sector; Reduce mandatory maternity leave from the current 410 days to a level closer to the ILO recommended minimum of 98 days. Slow school-to-work Pilot apprenticeships, internships and wage Carry out impact evaluations of the initiatives transition subsidy programs for young workers. and take successful measures to scale. Large number of youth who Taylor labor market activation measures are not in employment, to NEETs. education or training (NEETs) Low household savings Equalize the retirement age of men and women and increase retirement age beyond age 65; Incentivize participation in voluntary pension schemes; Improve financial education and information. Education Low participation in life-long Identify key barriers to achieving increased Establish a tripartite partnership between learning (LLL) participation in LLL. graduates, career centers and employers to develop effective education programs. High drop-out rates, Actively implement the Strategy for Prevention particular among Roma and and Reduction in the Share of Drop-outs and other disadvantaged groups Early Leavers Provide better access to early child-hood programs for Roma and other disadvantaged groups. Lack of good teachers in Promote competitive hiring of best teachers in Create salary structures that provide incentives under-served areas disadvantaged areas using available EU for quality teachers to move to under- structural funds. served areas. Low transversal skills in all End early selectivity and tracking of students Maintain general education through age 16 students into vocational education. before tracking into selective and vocation educational tracks. Other productivity-enhancing policies Maintain high productivity Improve business climate, by better enforcing growth and increase labor contracts, streamlining the issuance of construc- demand tion permits, providing faster access to electricity and improving bankruptcy laws; Improve infrastructure; Improve energy policies; Enhance innovation policies Savings Negative household savings Improve financial education and information Increase retirement age beyond age 65; Incentivize participation in voluntary pension schemes.. (continued on next page) xxiv | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Summary of policy recommendations (continued) Challenge Urgent policy actions Medium-term recommendations Improving Efficiency of Fiscal Spending Fiscal Management Increasing fiscal pressures Develop comprehensive long-term fiscal projections; Assess the long-term fiscal impact of policy changes. Health and Long-term Care Inefficient hospital-based Develop a hospital rationalization plan and a Implement the hospital rationalization plan and care Diagnostic-Related Group (DRG) pay- the DRG payment system; ment system. Convert redundant municipal hospitals into community centers that provide a whole range of LTC and rehabilitative services. Lack of alternatives to Improve capacity of primary care health Improve human resource policies to develop hospital-based care providers and provide incentives for care outside profession’s speciality mix that meets popula- of hospitals tion’s needs. Lack of financial protection Develop publicly provided LTC services with a Develop financial LTC insurance. from risk of becoming focus on community and home-based services. dependent Pensions Pension system is not Explore a small increase in contribution rates. Increase retirement age beyond age 65. self-sustaining Ad-hoc increases in pension Refrain from further ad-hoc pension increases. Improve long-term fiscal planning of the benefits pension system. Increased pressure to the Further strengthen the disability certifica- disability program as tion process. eligibility conditions of the old-age program are tightened Low pension coverage Consider an alternative financing mechanism to ensure equity, such as a universal pension financed from general revenues. Executive Summary | xxv Introduction After thousands of years of stagnant Bulgaria, this process was completed in just life expectancies, average longevity in 36 years. As a result, this major demograph- developed countries doubled within a ic change has been straining Bulgaria’s insti- century and societies around the world tutions, creating an urgent need for reforms. are grappling with the question what For Bulgaria, aging is not a prob- life in an aging society would be like. lem confined to the future. The old of Prima facie aging tends to be associated with the future are already entering the education frailty and deteriorating quality of life. Not system or the work-force. Current institu- surprisingly, the elderly in Eastern European tions and expectations influence their choic- countries tend to be dissatisfied with their es today and will affect economic growth. lives because many of their expectations Going forward, ever-smaller cohorts of have not been met (Deaton 2007). In richer younger people will follow. Transferring re- many countries, though, the elderly tend to sources from the young to the old without be more satisfied with their lives. Generally, burdening future generations with unsus- psychological studies show that as people tainable debt will become increasingly diffi- age, they tend to become happier as they fo- cult. Postponing reforms today is likely to cus more on priorities, become more open require more drastic responses in the future. to reconciliation, are less prone to crime and Moreover, the median voter will age with have less tolerance for injustice. Thus, if a the population, possibly making it more dif- country with an aging population can solve ficult in the future to garner public support the practical problems that older people face, for policy measures in support of the young it could be quite a happy place. or future generations. Many of the reforms Contrary to most countries in the discussed in this report would also benefit world with an aging population, Bul- today’s elderly. Old-age poverty in Bulgaria garia has become old before becoming is widespread and providing better public rich. Most developed countries with a high services could improve the quality of life of share of elderly had decades to adjust to the elderly. changes in the age structure. It took more This report is aimed at addressing than hundred years in France, nearly 70 the question of how the government years in the US and more than 50 years in can provide better public services for Hungary for the share of the older popula- today’s elderly while investing in Bul- tion (65 plus) to increase from 7 percent to garia’s future. It responds to a request of 14 percent (Gragnolati et al. 2011). In the Bulgarian Minister of Finance in xxvii November 2012 to assess the consequences vices and public finance. It identifies the that Bulgaria’s population aging and the de- main channels, the risks and the challenges cline in the size of the work-force will have the government is likely to face. It discusses on public services and finance. It is envi- the implications of different policy options sioned as the first part of a two-pronged and identifies reform opportunities. It starts study, with the second report focusing on by describing Bulgaria’s demographic trans- constraints to productivity growth at the formation and how it is likely to affect eco- macroeconomic and firm level. The decline nomic behavior and macro-economic in numbers of Bulgaria’s labor force and the outcomes (Chapter I). Labor-market and ed- changes in its composition could weaken the ucation policies are important to stem the country’s growth prospects. Boosting pro- decline in the labor force and support pro- ductivity will therefore be crucial to acceler- ductivity growth. They are discussed in ate income convergence. The discussion of Chapter II and III, respectively. Improving productivity-enhancing reforms through Bulgaria’s health-care system, analyzed in channels other than improving education Chapter IV, will be essential to address the and health will take place in the second health care demands of an aging population study. This study is also not intended to pro- in a sustainable and equitable manner. How vide an in-depth examination of the links to meet the long-term care needs of an aging between private savings and demographic population will be discussed in Chapter V. change, or of the impact of demographic The state of Bulgaria’s pension system and change on public revenues and the geo- options for reform are outlined in Chapter graphical implications of demographic VI. The sector analysis of the preceding change. These areas will require sepa- chapters will be combined in a coherent rate analyses. macro-model to simulate the impact of de- This report focuses on the impact mographic change on growth under differ- of demographic change on public ser- ent scenarios and shocks in Chapter VII. xxviii | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria CHAPTER I Demographic change, labor supply and the economic life-cycle Bulgaria is undergoing a profound so- 1990s, it climbed to 1.5 in 2011, placing it cio-economic transformation brought close to the European (EU) average. Previ- about by extraordinary demographic ously, it had declined steadily from 2.3 in change. Between 1950 and 1990, Bulgaria’s 1960 to 1.09 in 1997, the lowest TFR ever population grew from 7.3 million to 8.8 recorded for a European country in peace- million but then imploded, shrinking to 7.5 time (Vassilev 2005). Declining fertility million by 2010. The population is project- rates play an important role in population ed to fall to 5.9 million according to Euro- aging, making more recent cohorts smaller stat and to 5.5 million according to the than preceding ones and over time skewing projections according to the United Nations the age distribution toward the older Population Division by 2050. This decline (Gavrilova and Gavrilov 2009). The recent goes hand in hand with major shifts in the increase in the fertility rate will therefore age structure of the population, which are slow down the aging process in the medium driven by changes in fertility rates, mortali- term as the proportion of younger people in ty rates and migration. A graphic representa- the population swells. tion of Bulgaria’s population is a pyramid in Life expectancy at birth has im- 1950, a globe in 2010 and expected to be an proved only slightly since the 1960s and inverted pyramid by 2050 (Figure I-1). Bulgaria has fallen behind other new EU states. It remained fairly constant from the 1960s until 2000 and has climbed slowly Demographic trends and since then, going from 71.6 years in 2000 to projections 73.9 in 2011. This is one of the lowest life expectancies in the EU, where the 2011 av- Bulgaria’s demographic changes have erage was 79.7 years. Bulgaria’s low life ex- been driven by three key forces in the pectancy is driven by stubbornly high infant past: A low total fertility rate (TFR), a stag- mortality—the second-highest in the EU nant life expectancy prior to 2000 and large and twice the EU average - as well as a prev- emigration. After Bulgaria’s TFR fell to ex- alence of non-communicable diseases. Heart ceptionally low levels by the end of the attacks, heart failures, and strokes jointly are 1 Figure I-1: One century of demographic change in Bulgaria 1950 1990 100+ 100+ 95–99 95–99 90–94 90–94 85–89 85–89 80–84 80–84 75–79 75–79 70–74 70–74 65–69 65–69 60–64 60–64 Age Groups Age Groups 55–59 55–59 50–54 50–54 45–49 45–49 40–44 40–44 35–39 35–39 30–34 30–34 25–29 25–29 20–24 20–24 15–19 15–19 10–14 10–14 5–9 5–9 0–4 0–4 400 300 200 100 100 200 300 400 400 300 200 100 100 200 300 400 0 0 Population in Thousands 2010 2050 100+ 100+ 95–99 95–99 90–94 90–94 85–89 85–89 80–84 80–84 75–79 75–79 70–74 70–74 65–69 65–69 60–64 60–64 Age Groups Age Groups 55–59 55–59 50–54 50–54 45–49 45–49 40–44 40–44 35–39 35–39 30–34 30–34 25–29 25–29 20–24 20–24 15–19 15–19 10–14 10–14 5–9 5–9 0–4 0–4 400 300 200 100 100 200 300 400 400 300 200 100 100 200 300 400 0 0 Population in Thousands Female Male Source: Author’s calculations based on data from the United Nations Population Division (2011). the cause of 65 percent of all deaths in 1 Any increase in life expectancy will slow the popu- Bulgaria (Couffinhal 2012).1 lation decline but its impact on aging is ambiguous. Significant emigration contributed If life expectancy rises due to declining infant mor- to the steep population decline over the tality (or a more general decline in the mortality of people younger than the mean age of population) last two decades (Figure I-2). Since the be- population aging —similar to an increase in TFR— ginning of the 1990s, emigration represented will relent. If life expectancy rises due to a decrease a 6 percent loss of Bulgaria’s population and in mortality rates among older age groups, popula- 10 percent decline of the economically active tion aging will accelerate. 2 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure I-2: Determinants of population growtha a. Population Growth b. Determinants of Population Growth 8 18 1.1 0.4 15.3 0.2 3.7 5.6 14 2.5 6 7.1 5.0 5.0 10.4 3.8 3.5 10 9.4 4 16.4 15.2 15.1 11.4 10.9 6 2 2 0 –2 –2 –6 –10 –14 –11.6 –4 –2.6 –10 –7.3 –10.4 –10.4 –10.5 –9.6 –9.9 –6 –14 –1.6 –2.8 –1.4 –8 –18 Bulgaria EU10 EU15 EU25 Bulgaria EU10 EU15 EU25 Bulgaria EU10 EU15 EU25 Bulgaria EU10 EU15 EU25 1960–89 1990–2011 1960–89 1990–2011 Population Growth Net Migration Rate Mortality Rate Birth Rate Source: Author’s calculations based on Eurostat. a EU10 excludes Bulgaria. Statistics are un-weighted averages for EU11 and EU15 and weighted average for EU25 of annual population growth. population (OECD 2012a). Net emigration projected to be older than 65, while only alone explains about 17 percent of Bulgaria’s one in two will be of working-age. population decline since the 1990s. High un- As Bulgaria’s working-age popula- employment rates, declining living standards tion decreases, fewer workers will gen- during the 1990s and significant wage differ- erate the national income for the entire entials between Bulgaria and its neighboring population. Since the relative size of the la- countries have been identified as key reasons bor force is a key determinant of a country’s for emigration. Also, the lack of favorable income level, this decline is likely to depress employment opportunities and infrastruc- growth. Labor force participation rates ture for high-skilled people interested in de- in Bulgaria are low for some age groups and veloping their own business have been for women, so there may be some potential stressed as another reason for migration to stem the impact of a declining working- (Rangelova and Vladimirova 2004). age population on  labor supply by  increas- In recent years, demographics have ing, all things being equal, the labor force become more unfavorable as Bulgaria participation (LFP) rates among specific heads for the steepest decline in the population groups, assuming there is  suffi- working-age population world-wide by cient labor demand. The most effective way 2050.2 Until 2007, Bulgaria’s working age to stop the labor force from shrinking fur- population grew relative to the total popula- ther is however to stanch migration. In addi- tion and the largest share of Bulgaria’s popu- tion to  mitigating the decline in  the labor lation was of working age (Figure I-3). Since 2008, the working-age population as share of total population has started to decline. As 2Calculated as the percentage decline in population a result, there will be fewer Bulgarians of aged 15 to 65 using the medium variant of world-wide working-age to support an increasingly larg- population prospects from the United Nations Popula- tions Division (2011). Information regarding the meth- er share of dependent age groups (in this odology underlying the projections and the data is study, children under 15 and persons over available at http://esa.un.org/unpd/wpp/Excel-Data/ age 64,). By 2050, one in three Bulgarians is population.htm Demographic change, labor supply and the economic life-cycle | 3 Figure I-3: Change in age structure and dependency ratios 9,000 80 8,000 70 7,000 60 6,000 50 5,000 40 4,000 30 3,000 2,000 20 1,000 10 0 0 1950 1960 1970 1980 1990 2000 2010 2020 2030 2040 2050 1950 1960 1970 1980 1990 2000 2010 2020 2030 2040 2050 65+ 15-64 Less than 15 Student Dependency Ratio Old-Age Dependency Ratio Source: Author’s calculations based on data from the United Nations Populations Division (2011). force, increasing productivity will be crucial of  older workers and support extension to sustain future growth. of working lives. These policies will be dis- cussed in detail in other chapters of this report. The scenarios presented here do not fo- Labor force projections and cus on  one particular policy measure, but scenarios rather on  a  combination: increases in  LFP of  women, youth and, most importantly, In order to better understand the inter- older workers are part of each scenario, but action between Bulgaria’s demograph- the emphasis is different in each: ic changes and the evolution in the size of the labor force, a detailed analysis of • The baseline scenario (Scenario I) as- Labor Force Participation (LFP) rates sumes an increase in participation rates across age and gender was carried out. until 2020 in line with the ILO projec- The approach combines population projec- tions. The ILO projects LFP rates men tions from Eurostat and the United Nations (and women) aged 60–64 to climb with data from the ILO on current and fu- steeply from 41.5 percent (18.4 for ture LFP rates by gender and five-year age women) to 51.2 (23.3) percent, while groups to gauge the impact of the demo- LFP rates of young workers aged 25 and graphic transition on the size and structure 54 change only minimally by about of the labor force.3 ˜tage points between 2010 and 2020. As Labor force participation decisions a result of these assumptions, the ILO’s are influenced by many factors. Declin- LFP profile in 2020 mimics very closely ing fertility rates, for example, are associated a profile that would be projected if LFP with increase in female labor supply (Bloom profiles were to be simulated on the et al. 2010.) Public policies, such as pension basis of current legislated reform to in- reforms, tax treatment of second wage earn- ers or child care subsidies, can stimulate la- bor force participation, in  particular for 3For a discussion of differences in population projec- women. Education policy also affects labor tions for EU and United Nations, see Annex 1. ILO’s labor force participations rates are available at http:// force participation rates by age groups, gen- www.ilo.org/global/statistics-and-databases/lang-- der and education level. Moreover, health- en/index.htm. Population projections by education sector reform can help increase participation level are discussed in Chapter VII. 4 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure I-4: Labor force participation rates under different scenarios 100 90 Percentage of Population 80 70 60 50 40 30 20 10 0 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 2012 2050 - Scenario I 2050 - Scenario II 2050 - Scenario III Source: Authors’ calculations based on United Nations Population Division (2011). crease retirement ages of men and will drop by 39.6 percent between 2010 and women (Figure I-4).4 After 2020, LFP 2050, based on UN projections, and by 35.3 rates by gender and five-year age groups percent based on EU projections (Figure I-5). are assumed to remain constant under The steepest decline is observed among this scenario (Figure I-4). young workers during the 2010s and 2020s • The LFP convergence scenario (Sce- and among workers aged 40–64 between nario II) assumes that participation rates during the 2030s and 2040s. by age and gender converge by 2050 to The steep decline of the labor force levels that are currently observed in could be mitigated if labor force par- countries with the highest participation ticipation rates in Bulgaria increased rates among the 50+ population (Ice- over time. Bulgaria’s labor force is project- land, Norway, Sweden, and Switzer- ed to decline under all scenarios. Even under land). This scenario implies increases in the high LFP scenario (Scenario II in chap- LFP across the board, but especially for ter I), which assumes convergence of LFP women and older workers. For example, rates of younger workers to levels seen in LFP rates of all (female) 60 to 64 year Nordic countries and a significant increase old are assumed to be 77.8 (74.4) percent in LFP of older workers far beyond Bulgar- in 2050 under this scenario instead of ia’s current retirement age, the labor force 37.9 (23.3) percent under the baseline. would still decrease by 12.4 percent accord- • The maximum scenario (Scenario ing to the UN projections and 6.9 percent III) assumes a convergence in labor force according to Eurostat projections. The num- participation rates with high LFP coun- ber of younger workers would be increased tries, a  convergence of  female-to-male by a little more than 200,000 workers com- LFP profiles and an increase in the average pared to the baseline scenario if Bulgaria length of work life by 10 years by gender. were to reach participation levels as seen in The scenario implies LFP beyond 80 per- the Nordic countries. The biggest gain cent even among those aged 70 to 74 and can be interpreted as an upper bound. 4Bulgaria passed a law in January 2012, increasing the retirement age for men and women by 4 months to age 63 and 4 months for men and to age 60 and 4 months Under the baseline scenario, the for women. From 2013, retirement ages will continue work force is projected to fall steeply. to increase by 4 months each year, until reaching age According to UN projections, labor supply 65 for men in 2017 and age 63 for women in 2020. Demographic change, labor supply and the economic life-cycle | 5 Figure I-5: Population and labor force projections under different scenarios a. Population Projection b. Labor Force Projections by Scenario 100,000 60,000 50,000 40,000 0 20,000 –50,000 0 –100,000 –20,000 –150,000 –40,000 –200,000 –60,000 –250,000 –80,000 –300,000 Scenario I Scenario I Scenario II Scenario III –350,000 (UN Population (EU Population (LFP (Maximum) –400,000 Projections) Projections) Convergence) 2010s 2020s 2030s 2040s 2050s 15–39 40–64 65+ 15–39 40–64 65+ Source: Authors’ calculations based on United Nations Population Division (2011), Eurostat (2012) and ILO (2012). though, would come from an increase of can translate these life-cycle changes into workers beyond the age of 65: convincing macroeconomic outcomes. these older workers to stay in the labor force could potentially mobilize many hundreds of thousands of them. Of course, this would Demand for public services result in a radically different labor force, with a much higher share of older workers. Demand for key public services changes It would also require improvements in the over a  person’s lifetime. Demand for public provision of education and health health services and long-term care tend to in- care (as discussed in Chapters III and IV, re- crease with age. Older people tend to  con- spectively). Reducing net migration by an sume more health care as  illnesses, chronic average 60 percent a year, which is the aver- diseases and hospital visits become more fre- age difference between EU and UN popula- quent. Only at very old age does health care tion projections over a forty-year horizon, demand tend to decline. In Bulgaria, people would increase the labor force by about 8.1 in  their seventies have the largest public percent, more than under any of the LFP health care costs per person (see Figure I-3). scenarios except the combined maxi- Similarly, demand for long-term care tends mum scenario. to increase steeply after people turn 65 years of age (European Commission 2012a). Only the old tend to retire. As the population ages, The Economic Life-Cycle public expenditures on  health, long-term care and pensions can, thus, be  expected Population aging matters because indi- to increase. On the other hand, demand for vidual economic behavior changes with education tends to be concentrated at young- age. The decision to participate in the labor er ages. Younger people are also more likely force is  only one of  them. In  the course to commit crimes. Population aging is there- of their life people also change their demand fore expected to reduce public expenditures for specific public services, their savings be- for education as well as public order and safety. havior and skill acquisition. The expected in- Bulgaria’s current expenditure com- crease in the share of older people in Bulgaria position does not reflect the fact that it 6 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure I-6: Bulgaria’s hospital expenditure-age profile 3.5 (Hospital Costs per 40–49 3.0 2.5 Year Old = 1) Cost Index 2.0 1.5 1.0 0.5 0 0–9 10–19 20–29 30–39 40–49 50–59 60–69 70–79 80+ Source: Hospital Data from MoH. has the EU’s third-highest median age. Its spending of other EU countries—both in public expenditures on health and long-term terms of GDP and of public expenditures. care are among the lowest in the EU. In 2011, Bulgaria’s public health spending in terms of GDP was significantly lower than in other EU 5 EU10 refers to Bulgaria, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovak countries, amounting to 4.6 percent of GDP Republic and Slovenia. compared to 5.4 percent among the EU105 6 EU15 refers to Austria, Belgium, Denmark, Finland, and 7.4 percent among the EU15 countries France, Germany, Greece, Ireland, Italy, Luxembourg, (Figure I-7).6,7 Available data suggests that Bul- the Netherlands, Portugal, Sweden, Spain, and Unit- garia is among the European countries that ed Kingdom. 7 Public expenditures for health and long-term care as spend least on long-term care.8 At the same a share of GDP are even low when taking into account time, Bulgaria’s spending on defense, public that Bulgaria has the lowest income level of the EU. order and safety significantly exceeds the 8 See, Chapter V, for a discussion. Figure I-7: Bulgaria’s expenditure composition in comparison a. In Percent of GDP b. In Percent of Total Expenditures 60 100 9.7 10.2 12.9 11.6 11.2 50 13.2 80 6.4 6.0 5.7 11.3 10.6 9.2 7.4 3.1 13.9 13.7 40 5.3 10.8 12.6 7.0 12.1 5.4 2.9 60 19.4 10.5 3.8 5.2 11.8 9.1 3.6 3.1 6.4 13.7 12.9 30 4.5 3.5 5.7 3.8 5.4 4.1 13.3 14.8 14.3 7.6 4.8 40 12.9 12.8 3.9 5.5 7.4 12.9 20 4.2 5.3 6.7 10.5 5.2 5.4 4.6 5.0 4.1 20 38.1 39.0 10 20.0 36.2 31.3 34.1 14.4 17.4 27.2 10.7 12.9 12.4 0 0 2007 2011 2007 2011 2007 2011 2007 2011 2007 2011 2007 2011 Bulgaria EU10 EU15 Bulgaria EU10 EU15 Social Protection Health Economic A airs General Public Services Defense, Public Order and Safety Education Housing and Community Amenities Environment Protection Recreation, Culture and Religion Source: Eurostat. Demographic change, labor supply and the economic life-cycle | 7 Table I-1: Decomposition of public spending on secondary education Bulgaria EU27 2005 2006 2007 2008 2009 2010 2009 Secondary education spending 2.0 1.8 1.7 1.9 1.9 1.8 2.4 (in percent of GDP) Spending per secondary student 12.5 11.7 11.5 13.2 13.6 13.4 16.4 (in percent of GDP per WA adult) Gross enrollment rate 0.81 0.83 0.85 0.85 0.86 0.86 0.8 Dependency ratio 0.19 0.19 0.18 0.17 0.16 0.15 0.18 (School-age pop in percent of WA pop) Source: Authors’ calculations based on Eurostat (2012). Education spending has adjusted to to  be  short in  Eastern European countries. Bulgaria’s demographic change in re- The National Transfers Account (NTA)9 cent years. Bulgaria’s student-age popula- shows that Hungarians and Slovenians start tion has declined steeply over the last decade. saving on average at the age of 27 and be- In 2012, this population was only about 65 come net borrowers at 55. In Austria, Fin- percent of its 1995 level—the steepest fall in land, Germany, Spain and Sweden, people the EU. The government responded by im- start saving on average at 25 and stop saving plementing a per-capita school-funding around age  60. Hungarians and Slovenians model in 2007 and closing over 500 schools have therefore seven years less to save for re- in 2007 and 2008. Public education spending tirement than people who live in wealthier per student relative to GDP per capita in- European countries (Figure I-8). It  is  thus creased from 15.2 percent in 2001 to 20.5 not surprising that consumption at  older percent in 2009 for primary education and ages is  significantly lower in  Hungary and from 16.2 percent to 21.8 percent for sec- Slovenia. A  short savings period followed ondary education. Though secondary educa- by low consumption at old age is also likely tion spending, for example, declined in terms to apply to Bulgaria, given its relatively low of GDP between 2005 and 2010, spending effective retirement age of  63  for men per secondary student in terms of GDP per and  60  for women, low replacement rate working age adult and gross enrollment rates of  49.8  percent10 and the high incidence increased during this period (Table I-1). of  old-age poverty. More than  60  percent of  Bulgaria’s population aged  65  or  older is at risk of poverty or social exclusion. Savings People also change their savings be- havior with age. They tend to be net bor- 9 The National Transfers Account (NTA) measures all rowers when young and old and save during reallocations of income across age at the aggregate lev- el for many countries, but not Bulgaria (Mason and their working years. How long people will Lee 2010). The only EU10 countries in included in the save in the course of their life is influenced data base are Hungary and Slovenia. For more infor- by  many factors, including education sys- mation, see www.ntaccounts.org. 10 Gross income replacement rates show the level of tems, health, life-expectancy, employment pension income the first year after retirement as a per- opportunities, family values, public pro- centage of individual pre-taxed income (after employ- grams and individual wealth. The number er contributions, but including employee contributions) of  years during which people save tends at the moment of take-up of pensions. 8 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure I-8: Income and private consumption for selected countries 35 30 (In Thousands US$, PPP Adjusted) 25 Annual Flows per Capita 20 15 10 5 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90+ Europe Hungary Slovenia Source: www.ntaccounts.org. Note: Data for Europe is the un-weighted average for Austria, Finland, Germany, Spain and Sweden. Blue lines refer to income, red lines to consumption. The high incidence of old-age pov- 2005 (Figure I-9). Other European coun- erty in Bulgaria and the currently rap- tries, including countries in Eastern Europe, id aging of its population raise the have achieved positive household savings question how an acceptable standard of rates; for example, Hungary and Poland living for Bulgaria’s elderly can be have maintained average rates of 10.6 and funded. Options include delaying retire- 7.3 percent, respectively, for the period ment; relying on family or the community; 2005–10. Still, since the crisis Bulgaria’s increasing public transfers to the old or en- household savings have been steadily in- couraging savings. There seems to be scope creasing and the government has been im- to increase the legal retirement age beyond plementing some measures to encourage the recent legislated increases in retirement household savings. For example, gradual in- age (see Chapter VI) and to provide better creases to the retirement age have been in- employment opportunities for older workers troduced to reach 65 for men and 63 for (see Chapter II). Relying on family is likely women by 2017 and 2020, respectively. The to become increasingly costly for Bulgaria’s government has also put in place measures economy as the working-age population to increase the employment rate. Despite shrinks. Any increase in public transfers these actions, household savings are unlikely would need to be assessed very carefully in to increase significantly as long as unem- the context of a long-term fiscal framework ployment remains high, growth weak and since it could generate significant fiscal costs poverty wide-spread. as the population ages. As Bulgarians are expected to live longer and the proportion of older peo- Productivity growth ple increases, it will become impera- tive to encourage savings. In Bulgaria, As individuals age, they tend to be less household savings have been negative since able to perform certain tasks. There is Demographic change, labor supply and the economic life-cycle | 9 Figure I-9: Gross savings and sources of household income a. Gross Savings by Sector (Percent of GDP) b. Household Income by Source 30 100 20 80 10 60 Percent Percent 0 40 –10 20 –20 0 2005 2006 2007 2008 2009 2010 1999 2001 2003 2005 2007 2009 2011 Non-Financial Corporations Financial Corporations Others Entrepreneurship Other Earnings General Government Households Household Plot Pensions Wages and Salaries Source: Authors’ calculations based on National Statistical Institute. clear evidence that physical productivity de- As Bulgaria’s working-age popula- clines somewhat as people gets older. That tion declines, income convergence will mental productivity declines with age is how- require strong productivity growth. ever less clear cut. In fact, older workers tend Bulgaria’s high real annual real GDP per to outperform younger workers in positions capita of 5.8 percent between 2000 and 2008 involving management, communications and growth was sustained by impressive gains in conflict resolution skills. They fare worse in labor productivity. Labor productivity levels jobs requiring strong cognitive abilities. Older rose by 40 percent between 2000 and 2006 workers may also be at a disadvantage in more alone and its growth of 3.5 percent over the innovative sectors, since they tend to have last decade significantly exceeded that of the more difficulty adjusting to a rapidly changing EU. As a result, Bulgaria’s income relative to environment, though this could be due to the the EU countries (EU27) increased from fact that older workers tend to have stayed in around 28 percent in 2000 to 46 percent in the same workplace for a longer time rather 2011. However, despite the gains in income than age (Daveri and Maliranti 2007). convergence with the EU, Bulgaria’s GDP Several factors suggest that produc- per capita (PPS adjusted) of Euro 11,600 in tivity growth is likely to decline in the 2011 remains the lowest in the EU context of  Bulgaria’s demographic (Figure I-10). In the future, there will be change. First, there exists a strong link be- fewer workers available to generate the re- tween economies of scale, population densi- quired output to sustain per capita income. ty and technical progress. As  Bulgaria’s As a result, it has been estimated that pro- population shrinks, population density will ductivity growth rates of more than 4 per- decline and human capital externalities may cent per year—0.75 percentage points more weaken. Second, the decline and aging than in the past decade—will be required of the population may depress entrepreneur- over the next 30 years in order for Bulgaria ial activity and innovation. Third, there to converge to the income level of Portu- is a danger that aging reduces public savings gal—currently the poorest of the original as age-sensitive public expenditures increase EU member states (Mitra and Pouvelle 2012). and revenues decline, potentially crowding Improving Bulgaria’s business cli- out public investment. mate will be  important for increasing 10 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure I-10: Growth and convergence a. Average Annual Change in Income Gap 1990–2011 b. GDP per Capita, Percent of EU Average (EU27=100, in PPS) 2 100 84 80 1 80 73 66 65 67 62 58 60 Percent 45 49 0 40 80 71 –1 54 48 50 20 45 40 36 28 26 –2 0 1990–94 1995–99 2000–04 2005–08 2008–11 CZE HUN POL SVK SVN EST LTU LVA BGR ROM Real GDP per Capita Growth 2000–11 BGR EU10 PPP GDP per Capita 2000 (in % of EU27) Source: Author’s calculations based on WDI. Source: Eurostat. Note: Positive values indicate a decline in the income gap. Note: Date for Romania refer to 2000 and 2010. productivity growth.11 Improving labor the IT sector has seen a rebound. Data from market regulations, education, and health recent enterprise surveys indicates that Bul- services, which are discussed in detail in the garian firms that innovate tend to grow one- next chapters, can all contribute to produc- and-a-half times faster than non-innovative tivity growth.12 But reforms in  other areas firms. The latter also tend to  create more are also likely to be very important. Bulgar- jobs: in recent years, they enjoyed an 8 per- ia’s business environment, as  measured cent annual increase in employment, where- by the World Bank’s Doing Business index, as job growth in non-innovative companies was the second lowest among the EU10 coun- was stagnant. Improving R&D policies tries in 2012. The World Bank’s Doing Busi- in support of innovation will thus be impor- ness suggests that Bulgaria has much to gain tant (World Bank 2012b) and are core con- from making the business environment tributors to Bulgaria’s global competitiveness more predictable, for example, by  stream- in the Government’s National Reform Pro- lining the issuance of construction permits, gram 2011–2015. To meet the national tar- providing faster access to  electricity, im- get of  1.5  percent R&D/GPD by  2020, proving bankruptcy laws and enforcing con- Bulgaria would need to  greatly increase its tracts (World Bank 2012a). ability to absorb EU funds and boost public More and more effective innovation R&D spending. is needed for Bulgaria’s firms to climb Moreover, improvements in infra- up the value chain and stimulate the ex- structure and energy policies are likely ports of higher-technology goods. Bulgaria’s to significantly improve the perfor- exports are still dominated by  natural re- mance of Bulgaria’s firms. Contrary to source-based and unskilled, labor-intensive commodities. Its share of high-tech exports reached at 3.8 percent in 2012, slightly low- 11 As mentioned in the introduction the discussion of er than the EU10  average and substantially productivity-enhancing reforms through channels below the EU27  average of  15.6  percent.13 other than improving education and health will be rel- egated to a second study. But innovation in  Bulgaria has steeply de- 12 Also measures that reduce the size of the informal clined in many areas as a large number of sci- economy could contribute to productivity growth. entist and skilled workers emigrated, though 13 Data from Eurostat (2013). Demographic change, labor supply and the economic life-cycle | 11 other EU10 countries, Bulgaria has been slow young women and the elderly. The larg- to build highways and existing roads and the est gains will stem from increasing labor railways sector have suffered from years of force participation rates among the elderly. neglect (World Bank 2012c). Bulgaria’s econ- This will require considerable reforms to in- omy also remains high energy intensive de- centivize older workers to stay longer in the spite the closure of the most inefficient work-force and to retain marketable skills industrial plants early during transition (its during their lengthened work life. Such re- use of energy per € of GDP stands at four forms and incentives could include: further times the EU average). High energy depen- increases in the retirement age14; investment dency makes Bulgarian firms less competitive in life-long learning; and improvements in than others in Europe because they cannot the health care system.15 Investing in human adjust to energy related shocks. Recent em- capital can be a powerful tool for increasing pirical evidence suggests that improving in- the labor force and productivity. This is sim- frastructure and related policies, innovation ilar to the OECD (2012b) finding which and education could raise Bulgaria’s labor emphasized the importance of education productivity growth by one percentage point and labor productivity growth in compen- a year (Mitra and Pouvelle 2012). sating for the negative effects of aging on The next chapters will analyze economic development. in more detail the impact of Bulgaria’s Stemming the current trend of em- demographic change on  the different igration and attracting returning na- sectors of the economy. Bulgaria’s labor- tionals as  well as  new migrants will market challenges and opportunities for re- be important to slow down the work- form are discussed in  Chapter II. Since force decline. Policies that expand the la- education policies are important to stem the bor force would also be  likely to  stem the decline in the labor force and increase pro- exodus of  Bulgarian workers. Poorly func- ductivity, Chapter III provides an in-depth tioning labor markets, weak public-service analysis of Bulgaria’s education sector. How delivery, and an uncertain business environ- to address the health care demands of an ag- ment have all been found to  be  significant ing population in a sustainable and equitable determinants of  emigration in  Eastern Eu- manner is the focus of Chapter IV. Options ropean countries (World Bank 2007). Thus, for scaling up Bulgaria’s long-term care are policies in support of higher labor force par- discussed in  Chapter V. Bulgaria’s pension ticipation rates, such as  labor market poli- system and options for reform are lined out cies, education policies and health sector in  Chapter VI. These sector analyses feed reforms, which are discussed in detail in the into in a coherent macro-model that quanti- next chapters, if  combined with policies fies the impact of  demographic changes in favor of improving the business climate, on growth and fiscal variables under differ- could help increase Bulgaria’s labor supply ent policy scenarios outlined in Chapter VII. by increasing labor force participation rates at the same time as reducing emigration. Policy options 14 In January 2012, Bulgaria passed a law that increased the retirement age for men and women by four months Since Bulgaria’s LFP rates are current- from age 63 for men and age 60 for women. The law ly low for some age-groups, there is stipulates that from 2013 onwards the retirement age of some potential to moderate the decline men and women will increase by 4 months each year until reaching age 65 in 2017 for men and age 63 for of the labor force over the next 40 years women in 2020. by implementing reforms that increase 15 An extension of work-lives and increases in labor participation rates, in particular among productivity could also boost savings. 12 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria As Bulgaria’s population ages, its Bulgarian government has developed public institutions and policies will several strategy documents that discuss need to adjust. Bulgaria‘s population is ag- these topics. In 2012, the Council of Min- ing very rapidly. To address the expected in- isters adopted, for example, the “Updated crease in  age-related public spending, the National Strategy for Demographic Devel- government has several options: (i) prudent- opment in  the Republic of  Bulgar- ly reduce public services as they become ob- ia  2012–2030” which recognizes that the solete: the introduction of  per capita current demographic transition will pose se- spending in  education in  2007  combined rious challenges for public services and fi- with the policy of protected schools is a good nances. It  emphasizes the need to  promote example;16 (ii)  explore possible reductions active and productive aging. This aspect in  security-related public expenditures; is further developed in the “National Con- (iii) enhance productivity growth by invest- cept Paper for the Encouragement of Active ing in  human capital or  innovation and Life for the Elderly” and the “National Pro- by improving the business climate.17 gram for Active Life for the Elderly People” The government could also consid- which aim at  promoting the participation er further steps to promote savings of  elderly people in  the labor market, im- over the medium term by: i) increasing proving conditions for life-long learning the retirement age to reduce the length of and supporting healthy aging. These chal- the life-cycle deficit and reduce the pension- lenges will be  discussed in  detail in  the fund deficit;18 ii) slightly increasing contri- next chapters. bution rates (see Chapter VI); iii) encouraging supplementary voluntary pension insurance (Pillar II) by increasing the existing tax re- 16 See Chapter III for a detailed discussion. lief of up to 10 percent of pensionable in- 17 The World Bank’s Doing Business ranked Bulgaria come from personal income taxes; and 66 in 2013 (World Bank 2012a) with respect to its iii) encourage financial education to help business climate, suggesting significant scope for im- provements, particularly, in areas such as dealing with people make better financial choices. construction permits, getting electricity and trading Aware of  the economic and social across borders. challenges of demographic change, the 18 See Chapter VI for a detailed discussion. Demographic change, labor supply and the economic life-cycle | 13 CHAPTER II Labor market policies to mitigate the impact of population aging The sharp decline in Bulgaria’s labor population, one of the highest rates in supply, as a result of projected popula- the EU. tion trends, could be mitigated if labor The Government of Bulgaria has put force participation (LFP) were to in- forward an  ambitious reform agenda crease. In Bulgaria, the LFP rate, which is to improve labor market outcomes. It has the sum of the employment and unemploy- set clear development targets to  be  achieved ment rate, is low for specific population by 2020 in the areas of labor market, education groups, such as the youth and older people and poverty reduction, aligning its National (young and elderly women, in particular) as Reform Programme (2011–2015) with the well as the Roma. As mentioned in the pre- Europe 2020 Strategy. By 2020, Bulgaria aims vious chapter, increasing labor force partici- to  achieve an  ambitious employment rate pation rates in Bulgaria could reduce the of  76  percent, up  from the  65  percent rate projected decline significantly. in  2010, which was below the EU  average. The current labor market situation Within the EU, only two the Netherlands and is challenging, with relatively low em- Sweden had a  higher employment rate ployment rates and stubbornly high than  76  percent in  2010. Since employment unemployment rates. In 2010, Bulgaria’s rates among middle-aged men are already employment rate was lower than the EU av- high in  Bulgaria, any strategy to  achieve this erage. The employment rate among 20–64 objective would need to focus on youth, wom- year-olds was 65.4 percent, compared to en, older people and the Roma population. 68.5 percent for the EU-27 (Figure II-1). This chapter explores labor market Unemployment rates, which are generally reforms that could help boost employ- more volatile than employment rates, were ment. Job creation requires a combination 10.3 percent in 2010, up from a low of 5.6 of demand side and supply side measures to percent in 2008, prior to the onset of the work in tandem. Policy areas for job cre- Great Recession in 2009 (World Bank ation include: (i) promoting macroeconomic 2011a). They have risen to 12.3 percent in stability; (ii) improving the investment cli- 2012. Youth unemployment in Bulgaria was mate; (iii) refining labor market regulations; 28.1 percent in 2012, above the European (iv) designing an employment-friendly tax average of 22.8 percent. Long-term unem- system and (v) developing an adaptable ployment was 6.8 percent in terms of active workforce through the education system. 15 Figure II-1: Employment rates for 20–64 year-olds in 2010 80 70 60 50 40 30 20 10 0 United Kingdom Malta Hungary Italy Spain Romania Greece Lithuania Ireland Poland Slovakia Latvia Bulgaria Estonia Belgium France Slovenia Portugal Luxembourg Finland Germany Austria Cyprus Denmark Netherlands Sweden EU (27 countries) Czech Republic Source: Eurostat. Note: Employment rates are for 20–64 year-olds. Given the challenges Bulgaria is facing, all of GDP for the EU area. NEETs have usual- five policy areas are important. However, ly dropped out of school without qualifica- given the scope of this report, this chapter tions and are likely to  be  dependent focuses on (iii) and (iv) and Chapter III on on welfare programs. Although NEETs are (v) and does not aim to assess how labor de- heterogeneous, Eurofound (2012) finds that mand in general could be strengthened in the largest group of NEETs tends to be the Bulgaria. This analysis is based on the prem- unemployed. But NEETs also include other ise that job growth in Bulgaria can be fos- vulnerable subgroups such as  the sick and tered by general improvements to work disabled; and young care-givers. NEETs im- incentives and by removing barriers to em- pose a significant cost on the economy and ployment faced by groups identified as hav- even on society. For instance, NEET youth ing low labor force participation rates: the are less likely to trust other members of soci- youth, older people, women, and the Roma. ety and to  participate in  civic and political activities (World Bank 2012e). Employment rates among older Labor Force Participation Bulgarians are relatively low. The em- by Population Groups ployment rate for 50–64 year-olds is 54.5 percent in Bulgaria, more than 10 percent- The youth in Bulgaria are more disen- age points lower than the EU average and gaged from employment and educa- significantly lower than that of top perform- tion than in any other EU-27 country. ers (Figure II-3). A number of factors ex- Bulgaria has the highest rate of young peo- plain this, including a lack of gradual ple who are counted as not being in employ- retirement options; lack of training oppor- ment, education, or training (NEET) in the tunities for elderly workers; and negative EU. In fact, more than one out of every five perceptions held by employers of older young Bulgarians is  classified as  NEET workers and their health problems (World (Figure II-2). Eurofound (2012) estimates Bank 2013).19 In addition, it is believed that that the lack of  labor market participation by NEETs costs Bulgaria about 2.3 percent 19Similar patterns can be observed in other transi- of  GDP, nearly double the  1.2  percent tion countries. 16 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure II-2: NEET rate among 15–24 year-olds in 2009 25 20 15 10 0 Czech Republic Bulgaria Italy Greece Spain Ireland Romania Cyprus Latvia Hungary Portugal Slovakia Estonia Belgium France Poland Lithuania Malta Slovenia Finland Sweden Germany Denmark Austria Luxembourg Netherlands EU (27 countries) United Kingdom Source: World Bank staff calculations using Eurostat data. factors such as ill-adapted workplaces, nega- structure meant to induce it. Workers who tive stereotypes of older workers and social experienced age discrimination on the job norms about work at older ages also deter are generally more likely to separate from older workers from participating in the la- their employers than workers who did not bor market. ( Johnson and Neumark 1997).20 Opinion surveys show discrimina- Employment rates are lower for tion against workers older than 55 is women relative to men. This gap is par- perceived to be widespread in Bulgaria ticularly pronounced for young and elderly (Figure II-4). More than two-thirds of Bul- women. While employment rates for women garians believe that people over 55 face dis- crimination in the labor market (European Commission 2012c). The perception of age 20The EU enacted specific legislation requiring Mem- ber States to prohibit age discrimination in employ- discrimination can erode older workers’ ment. All member countries have now legislated to commitment to their employer, which in make it unlawful to discriminate in the labor market turn hurts productivity and the incentive on grounds of age (or other factors). Figure II-3: Employment rates among 50–64 year-olds 80 70 60 50 40 30 20 10 0 Hungary Czech Republic Sweden Germany Denmark Estonia Netherlands Finland Lithuania Latvia Cyprus Austria France Ireland Luxembourg Portugal Slovakia Bulgaria Belgium Spain Italy Romania Poland Slovenia Greece Malta United Kingdom EU (27 countries) Source: World Bank staff calculations using Eurostat data. Labor market policies to mitigate the impact of population aging | 17 Figure II-4: Discrimination perceptionsa 80 67 69 62 66 66 66 61 60 55 50 50 40 20 0 Poland Romania Lithuania Latvia Slovak Czech Estonia Slovenia Bulgaria Hungary Republic Republic Source: World Bank staff calculations using European Commission (2012b). a Y-axis shows percentage of respondents who believe that age above 55 places the worker at a disadvantage for accessing a job between the ages of 30 and 54 are above 76 women may be childcare provided by grand- percent, which is relatively high, they are be- mothers. Increasing women’s access to eco- low 22 percent among young women, aged nomic opportunities can have positive 15 to 24 in 2009 (Figure II-5). Employment impacts on children’s health, education and rates among young women are also signifi- nutrition, thereby affecting not only current cantly below the European average of 32.7 but also future economic growth (World percent. The likely cause of the delayed entry Bank 2011b). by women into employment is marriage and Bulgaria is among the countries in child care.21 In addition, only 63 percent of the world with the longest mandatory 55–59 year-old women is employed com- paid maternity leave. Mandatory mater- pared to 72 percent for men. The ratio drops nity leave is 410 calendar days, the longest sharply to 16 percent for 60–64 year-old women while 42 percent of men in that age 21An additional factor is that enrollment rates of wom- bracket are still employed. One reason for en in tertiary education are also significantly higher the low employment rates among older than enrollment rates of men. Figure II-5: Labor market rates for Bulgarian women and men in 2009 (in percent) 90 90 70 70 50 Women 50 Men 30 30 10 10 –10 –10 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 Neet Women Inactive Women Neet Men Inactive Men Unemployed Women Employed Women Unemployed Men Employed Men Source: World Bank staff calculations using EU-SILC data. Note: Y-axis is in percent and x-axis represents age groupings. 18 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria among 141 countries (World Bank 2012f ).22 aggravated where wage gaps between men The average paid maternity leave period for and women are significant to  begin with, Eastern Europe and Central Asia is 185 days as is the case in Bulgaria. Women in Bulgar- and for OECD countries 109 days. During ia earn on average only 59 percent of what maternity leave, women in Bulgaria receive men earn, putting the country in 98th place 90 percent of their current wage.23 In addi- in  the ranking on  Wage Equality (World tion to the paid maternity leave, Bulgaria Economic Forum 2012).26 grants 320 days of mandatory paid parental Roma already constitute between 9 leave which can be taken by either the father and 19 percent of new labor market en- or the mother.24 This can prolong the effec- trants in Bulgaria,27 yet employment tive maternity leave to a total of 730 days if rates among Roma are very low. In the woman chooses to take the parental 2011, only 42 percent of Roma men report- leave as well.25 ed working (any kind of employment for While providing maternity leave pay, including informal employment), while is important, long maternity leave pe- only 26 percent of Roma women report the riods can have unintended adverse con- same employment status (UNDP/WB/EC sequences for women. For example, 2011). This is significantly less than the 63 maternity, paternity, and parental leave ar- percent and 56 percent, respectively, of men rangements affect the choices that women and women in the general population make and the opportunities available to them (Figure II-6). It is also interesting to note in the working world. Introduced to protect that employment rates among non-Roma women, such laws ultimately limit their neighbors, while lower than that of the gen- choices and can adversely affect their ability eral population, are still considerably higher to find employment. A good mix of mater- than the Roma employment rates. These nity, paternity, and parental leave is needed low employment rates tend to be deter- so  that employers do  not have incentives mined by factors that are well defined by the to  avoid hiring women. There appears time an adult worker starts seeking employ- to be no ideal mix of leave. Different gov- ment, such as the education level. For exam- ernments favor different options to suit their ple, when education is constant, the Roma particular social and economic require- ments, and the benefits thus vary widely 22 45 days of the maternity leave must be taken before by  region. One policy to  consider would the child’s expected birth date. 23 Provided that that the woman has worked for 12 be to shorten the mandatory maternity leave months prior to taking maternity leave and and increase child-care options instead. has insurance. Another example of  well-intended 24 Parental leave payments are equal to the mini- gender policies with potentially ad- mum salary. 25 Just 48 out of 141 economies offer parental leave, and verse effects on women are differential in only 34 is any benefit paid. Parental leave is more retirement ages. In  Bulgaria the retire- common in middle and high-income economies. ment age for women is currently 60, 3 years Since 2009, fathers in Bulgaria are entitled to a 15 day below that of men. Difference in retirement period of leave on the birth of a child. They are also permitted to use the remaining portion of the 410-day age can create disparities in  lifetime earn- maternity leave when the child is between six and ings, pension benefits and retirement savings twelve months old. Bulgaria has also a special scheme (Levine, Mitchell and Phillips 1999). In ad- entitling grandparents to take parental leave. dition, early retirement may result in wom- 26 Other countries in Eastern Europe and Central Asia en not being promoted to senior management fare significantly better. For example, in Macedonia, Georgia, the Kyrgyz Republic, Tajikistan and Kazakh- positions, thus providing men with better stan the female to male wage ratio stands at 77 percent. career promotion opportunities (Ad- 27 The difference is explained by the methodology used ams 2002). Life-time earning differences are to count the number of Roma. Labor market policies to mitigate the impact of population aging | 19 Figure II-6: Employment rates for reduce underreporting, it  bears the danger Roma and non-Roma neighbors of  pricing out low-wage earners from the formal labor market. Since a priori it is not 80 clear whether the costs of introducing a pro- 60 gressivity element in labour taxation in terms Percent of underreporting outweighs potential ben- 40 efits of  increasing labor force participation of  low-wage earners in  the formal sector, 20 any change to the tax system would require 0 detailed and comprehensive analysis. General Non-Roma Roma The government has implemented Population Neighbors a  number of  policy reforms in  recent Men Women years to  address these labor-market Source: General Population: Eurostat 2011 Q2; Non-Roma challenges. These include, but are not lim- neighbors and Roma: UNDP/WB/EC regional Roma survey (2011). The latter survey is based on a random ited to: improving the quality of  employ- sample of primary sampling units with proportion Roma. ment services provided by the Labor Offices Employment rates are for 15–64 year old. Directorate (LOD); improving the effective- ness of  measures and programs undertaken by  the Employment Agency (EA); easing and non-Roma neighbor drops from 21 to school-to-work transitions with the “New 8.2 percent for men. Similarly, for women, Start” program; launching an integrated sys- the Roma and non-Roma neighbor drops tem for labor market demand and supply and from 20 to 11.4 percent.28 Still, some labor developing a unified national vacancy por- market constraints have been identified by tal; reducing the share of  undeclared work researchers, including a high tax burden on through increased inspections; and improv- labor and a weak job-search system (Bogda- ing labor market flexibility and job security nov and Angelov 2006). (Republic of Bulgaria 2011). Labor taxes are less progressive in Bulgaria than in any other European or  OECD country. Bulgaria implements Policy Options a flat tax that is unrelated to the earning lev- el of  the individual. However, unlike Providing affordable alternatives for the in  many countries that also implement flat care of children and the elderly could taxes, income tax credits and other means boost female employment by reducing of increasing progressivity are absent in Bul- the opportunity cost of working outside the garia. As Figure III-7 shows, the gap in the home. In 2011, the Slovak Republic experi- tax wedge between higher wage and lower mented with changes to maternity benefits, wage earners is negligible in Bulgaria, while allowing benefits to be combined with work in  all other European or  OECD countries and offering slightly higher pay to partly off- there is some measure of progressivity em- bedded into the system and hence the gap 28 To investigate whether employment rates are low is  in  the tax wedges is  positive for other when education is held constant, we run separate re- countries. Lack of progressivity in labor tax- gressions for men and women with data for all work- ation can be  particularly harmful to  low- ing age people and control for age, age square, wage workers, who in  turn tend to  be  the education levels, and community fixed effects (thus ef- fectively comparing Roma and non-Roma neighbors younger and older workers, as well as wom- in the same community). en and ethnic minorities or part-time work- 29 See Betchermand, Daysal and Pages (2008) and ers.29 While lack of  progressivity is  likely Koettl and Weber (2012). 20 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure II-7: Gap in tax wedge between high and low wage earners in 2008 20 Ireland Belgium France Switzerland Austria Germany 15 Progressivity of Tax Wedge Italy Hungary Netherlands (Percentage Points) United Kingom Spain Slovenia 10 Finland Norway Portugal Czech Republic United States Slovak Republic Greece Latvia Macedonia Lithuania Sweden Estonia 5 Japan Poland Romania Bosnia - Federation Bosnia - Republika Srpska Serbia Bulgaria 0 0 5 10 15 20 25 30 35 40 45 Tax Wedge for Single at 33 Percent of Averge Wage Source: Koettl and Weber (2012), based on OECD Tax and Benefit Models. Note: The tax wedge measures personal income tax and social security contributions paid by workers and employers as a share of total labor costs. Tax wedge for a single person with no children at 33 percent of average wage versus progressivity of tax wedge. Progressivity refers to the increase of the tax wedge between 33 and 100 percent of average wage, in percentage points. set child-care costs.30 With sufficient child- programs involving apprenticeships, intern- care supply, the reform is expected to increase ships, and job subsidy programs for young women’s labor force participation (World people to encourage them to take and keep Bank forthcoming). Bulgaria could also con- jobs. The success of these measures would re- sider reducing its very generous maternity quire contributions and partnerships between leave provision and use the savings to subsi- the Government, employers and trade unions. dize child-care options. In addition to piloting these initiatives, it Promoting flexible work arrange- would be important to undertake rigorous ments may draw more youth, older impact evaluations before any scale-up is con- workers and women into employment. sidered. Similar programs have been imple- For youth, it may be crucial to be able to mented in countries around the world and combine work with studies. For others, it some of them, especially in Latin America may be important to work part-time or at where youth unemployment tends to be ex- home to facilitate child-care duties. For old- ceptionally high, have passed rigorous evalua- er workers, expanding part-time and home- tions. Chile, Peru, Colombia and other based work may encourage individuals to countries can offer design options to Bulgaria. remain in the labor market until the stipulat- ed retirement age.31 30 Although the reform has not been evaluated yet, it is New apprenticeship, internship and expected that the reform is expected to increase wom- wage-subsidy programs for young work- en’s labor force participation, as opposed to subsidizing ers could be piloted and evaluated. Given women to stay at home. 31 Qualitative work in Croatia and Poland found that the particularly acute problem of youth half of those aged between 55 and 70 would have liked NEETs in Bulgaria, the school-to-work tran- to delay retirement, but that appropriate arrangements, sition might be improved by designing pilot such as part-time contracts, were not always available. Labor market policies to mitigate the impact of population aging | 21 An effective youth NEET strategy is mixed, so the measures need to be careful- needs to retaining young people in for- ly evaluated and, if brought to scale, mal education and training and be monitored. well-targeted. Bulgaria could scale up such Targeted, early childhood programs measures as job placement services, training may be  an  effective way of  improving and remedial education for older youth and employment among the Roma over back-to-school programs, enabling young time. Pre-schooling has been proven to im- people to acquire skills that are valued by prove the later scholastic performance of lin- employers. It could also entail outsourcing guistic minorities, including the Roma.While the full range of activation services to quali- the general Bulgarian population is declining fied agencies, for example to agencies with rapidly, the Roma population is  increasing experience in working with disadvantaged and it  is  important that they find employ- youth, such as the socially excluded Roma. ment, especially among young Roma. This The experience from other countries sug- will be  discussed in  more detail in  the gests that the impact of activation measures next chapter. 22 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria CHAPTER III The role of education policies As discussed in  the previous chapter, they are able to reap the returns to education groups with low LFP rates is only one over a longer period of time (Becker, Mur- of many reasons why investing in edu- phy, Tamura 1990).33 As Bulgaria’s popula- cation will be essential to sustain Bul- tion is projected to enjoy an increasingly garia’s long-term growth. As mentioned longer life expectancy at birth (see Chapter I), in Chapter I, reducing the convergence gap investment in human capital is likely to in- between Bulgaria and the rest of the EU will crease. Bulgaria’s future generations should require sustained improvements in produc- thus be more educated than today’s. tivity and a shift to better-paid economic ac- There is  significant scope for har- tivities generated by employees with higher nessing education to  mitigate the im- and better skills. Education increases pro- pact of demographic change. Bulgaria’s ductivity, notably by  fostering innovation. student population has declined steeply The urgency of focusing more on education in  recent years. The government has re- has been heightened by  the emigration sponded by moving to a per-capita funding of 50,000 scientists and skilled workers since model and consolidating schools. It has also the  1990s (Mansoor and Quillin, 2007). implemented programs to help young peo- Moreover, technology changes dictate that ple get jobs, reduce the number of drop-outs far-sighted companies need to invest in the and improve teacher training. Still, many continuing education of  their employees more reform opportunities exist. Promising or  fall behind—and better-educated em- ones include enhancing the quality of edu- ployees are easier to  retrain. Finally, there cation across all sectors, promoting lifelong is strong empirical evidence that the better learning and skill-building at all age levels, educated are healthier and live longer (Cut- focusing on under-enrolled sectors, expand- ler and Lleras-Muney  2006; OECD  2006). ing overall monitoring and evaluation, pro- All of these advantages of education are only moting evidence-based policymaking, and realized, however, when the education examining efficiencies at all levels of educa- is of good quality.32 tion, to ensure the best use of available funds. Bulgaria’s demographic change is likely to increases investment in educa- tion. Human capital theory predicts that in 32 Education may also affect demographic change by reducing fertility (Basu 2002). societies with higher life expectancy at birth, 33 Education in turn is likely to affect all three funda- all other things being equal, families have mental demographic components, fertility, mortality incentives to invest more in education as and migration. 23 Figure III-1: Public education expenditures in percent of GDP 10 8.7 8 7.2 7.3 6.5 6.8 6.6 6.0 5.9 5.7 5.9 5.8 5.7 6.1 6 5.1 4.1 4.4 5.1 5.6 5.6 5.1 4.7 5.0 4.2 4.6 4 2 0 NOR IRL DNK FIN GBR SWE LUX BEL AUT NLD DEU FRA ITA ESP GRC PRT SVN HUN SVK CZE POL EST LVA LTU ROM BGR EFTA EU15 EU15 EU15 EU10 EU10 EU10 North Continental South Continental North South Source: Authors’ calculations based on Eurostat. Bulgaria’s basic education in the only about inefficiency of public spending, context of Demographic change but also about deteriorating service quality as resources were spread too thinly across a Bulgaria’s public education spending in large number of small schools. The govern- terms of GDP is low, but it has been ment responded by implementing a per-cap- catching up with the EU average. Bul- ita school funding model in 2007 and closed garia’s education spending has increased over over 500 schools in 2007 and 2008 alone and the last decade from 3.88 percent in 2000 to 119 schools between 2009 and 2012 4.6 percent in 2009 (Figure III-1). At the (Figure III-2b). same, the student population declined three To ensure that access to education times faster than it did in the EU. The de- in remote villages was not reduced by cline started in the mid-1990s and accelerat- consolidating schools, the Government ed in the new millennium. Between 2000 introduced the notion of “protected and 2010, the primary and secondary stu- schools,” which were kept open so that dent population plummeted by 40 percent children need not travel excessive dis- (Figure III-2a). As a result, public education tances (generally over 30 km). Such dis- spending per student relative to GDP per tances were found to contribute to high capita has increased steeply from 15.2 per- drop-out rates (World Bank, 2010). Cur- cent in 2001 to 20.5 percent in 2009 for pri- rently, 127 schools around the country have mary education and from 16.2 percent to been formally recognized as protected. 21.8 percent for secondary education. Bul- Meanwhile, better-attended and/or newly garia’s overall public education spending per built schools became hubs and were desig- student relative to GDP of 27.1 percent nated “focal schools.” They attracted some slightly exceeded the EU average of 27 per- of the savings of the consolidation process cent in 2009. Still, primary and secondary and their standards were raised. As noted in public education spending per student rela- the Strategy for Prevention of Early School tive to GDP per capita were still below the EU average in 2009.34 The Bulgarian government has re- 34In 2009, primary, secondary and tertiary public educa- sponded judiciously to the steep decline tion spending per student relative to GDP per capita in Bulgaria amounted to 20.5 percent, 21.8 percent and in its school-age population. The result- 48.3 percent, respectively. The EU average in 2009 was ing diminished enrolments and overcapacity 22.6 percent for primary, 26.6 percent for secondary and of schools and staffing raised concerns not 40.8 percent for tertiary education spending. 24 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure III-2: Evolution of student numbers and ratios a. Students in Percent of Population b. Student Ratios 110 310 13.7 100 300 13.4 90 290 13.1 80 280 12.8 70 270 12.5 60 260 12.2 1995 1997 1999 2001 2003 2005 2007 2009 2011 2000–01 2001–02 2002–03 2003–04 2004–05 2005–06 2006–07 2007–08 2008–09 2009–10 2010–11 2011–12 Bulgaria EU10 EU25 EU15 Student/School Ratio Student/Teacher Ratio Source: Authors’ calculations; Eurostat. Source: Authors’ calculations based on data from Bulgarian NSI. Note: Index (1995=100). Leavers (2013), drop-out rates improved discontinued in many schools, to free up the dramatically at protected schools since they required physical infrastructure (classrooms for were first introduced in 2007/8. the afternoon part of the full day schooling). In cities, however, student popula- This requires expenditures to build or reno- tions are increasing and struggling to re- vate schools and to hire more and better-qual- ceive basic education. Migration flows from ified teachers. In these cases, expenditures small towns and villages to the large cities have across the education sector will need to in- strained their infrastructure. Several cities, in- crease to promote optimal student outcomes. cluding Sofia, have been forced to develop Indeed, significant resources (over €70 mil- two-shift school days. Such schedules are lion) from EU structural funds have been allo- known to lower education quality, since full- cated for teacher qualification and training, in time schooling with after-school tutoring and addition to the funds provided by the state classes are among the most effective mecha- budget, to meet the expanding needs of the nisms to reduce early school-leaving.To return education sector. to full-day education for all grades (from 1 to Quality of Bulgaria’s basic educa- 7) in all schools in Bulgaria in the coming tion has been declining in recent years. years would require that the double shifts be Bulgaria’s score on the 2011 Progress in Table III-1: Implications of demographic challenges for Bulgaria’s education system Response Data/Activities Primary/Secondary School Closures Total number of closures: 682 yy2005–2009: 563 yy2009–2012: 119 Access/quality concerns Protected Schools Identified: yy2011/2012–119 yy2012/2013–127 Bulgarian University Ranking System (BURS) developed to provide access to data on programmatic quality across the tertiary higher education sector Teacher training/recruitment shifts Improved teacher training for student retention, particularly of Roma students Source: MEYS. The role of education policies | 25 International Literacy Study (PIRLS)35, school reforms, the results corroborate that released in 2012, showed a decline com- Bulgaria’s education reform has been on the pared to the country’s own results in right track (World Bank 2010b). previous PIRLS rounds. A mirror find- ing of this outcome in the 2011 PIRLS is that around 40 percent of the Program for Tertiary education and International Students Assessments (PISA)36 demographic change 2009 student cohort was classified as “func- tionally illiterate,” unable to read well At the university level, the number of enough to extract and summarize relevant institutions as well as their enrollment information from text documents (World expanded (Table III-2) over the last de- Bank 2013). Improving quality of education cade. As the number of secondary students in Bulgaria is also hampered by the fact that dropped, the number of tertiary education the variance of scores on PISA is exception- students increased by 16 percent. The net ally large between schools. This inequality enrollment rate for tertiary education grew between schools seems to be driven, at least from 26 to 42 percent of the population in part, by the early selection of students into aged 19–23 between 2000 and 2012. This general and vocational tracks and into ‘elite’ increase was driven by the emergence of and ‘ordinary’ school. Addressing these in- new private universities as the public ones equalities will be critical to improving the expanded. Today, Bulgaria has 51 Higher economic chances of disadvantaged groups Education Institutions (HEIs), of which 18 and the overall quality of education in are private and 33 are public. Though ter- Bulgaria. tiary attainment rates across the population Even with this diminished quality, have increased, financial and social returns however, Bulgaria still ranked compar- to tertiary education attainment remain atively high in PIRLS 2011 (22nd out of high in Bulgaria and unemployment among 45 participating countries) in terms of students with tertiary education low. average performance of primary edu- Despite quantitative and structural cation outperforming France, Poland, achievements during the past two de- Spain, Slovenia and Romania. Bulgaria fea- cades, higher education in Bulgaria tured also among the 10 countries with continues to face challenges with regard highest share of students reaching the high- est PIRLS benchmarks, outperforming most OECD countries. Moreover, in terms of 35 The PIRLS is an international comparative study of performance gains in PISA scores in 2009 the reading literacy of young students (4th-grade stu- over 2006, Bulgaria ranks 7th, i.e. among dents in the United States and students in the equiva- the top improvers worldwide, with only two lent of 4th grade in other participating countries).. 36 PISA is a global assessment tool administered by the regional peers—Serbia and Romania— OECD that aims to evaluate education systems world- showing higher gains (41 and 28 points). wide every three years by assessing 15-year-olds’ com- Though there is no clear evidence these ris- petencies in the key subjects: reading, mathematics ing test scores were directly related to the and science. Table III-2: Growth of enrollments in tertiary education in Bulgaria Academic year 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 Growth in total 4 percent 2 percent 7 percent 2 percent 4 percent 5 percent –1 percent enrollments Source: National Statistics Institute. 26 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria to quality, efficiency and accountability. rates of close to 50 percent, however, the ter- The Ministry of Education,Youth and Social tiary system can no longer rely on growth Sciences (MEYS) is committed to compre- from traditional student populations. Re- hensive reforms across the sector, continued forms to the sector, including potential clo- innovations and improvements could further sures, mergers and the redeployment of improve it. The current funding model is institutions to serve changed academic goals based on a per-capita formula that takes into must be considered to stabilize the sector in account enrollments, performance and com- the face of diminished enrollments. petitive research funding needs. The Ministry The focus of learning must be- is continuing to seek new mechanisms for come more adaptive. The growing de- improving the quality of education, including mand for generic cognitive/non-cognitive greater use of performance-based contracts. skills requires sufficient and targeted re- A new strategic report (World Bank 2012g) sponses from education and training sys- for comprehensive reforms of the tertiary sys- tems, particularly in the training of teachers tem proposes various mechanisms—includ- and the development of adaptive curricula. ing financing formula recalibrations, quality It is not enough to provide students with assurance system improvements, and gover- learning modules geared towards memo- nance changes—to improve the outcomes of rizing facts. Employers across the world higher education investments so that quality cite the need for workers who can learn and efficiency become even stronger drivers quickly and work well in adaptive team en- in an autonomous yet accountable higher ed- vironments. Youth do better in acquiring ucation sector. skills, but many often acquire the wrong Responding to these challenges, the set (both generic/transversal and techni- government implemented the Bulgarian cal). Bulgaria has the tool in place (the Universities Ranking System (BURS) in BURS) to access powerful data on what is 2010. In use since 2010, the BURS provides and is not working in tertiary education users with a tool to compare study programs programs as it relates to outcomes. In addi- offered by Bulgarian tertiary education institu- tion, and with World Bank support, fur- tions against a set of 51 indicators. It provides ther information about this balance of readily available, hard and objective data that cognitive and non-cognitive skills in the complement the institutional and program workforce is being gathered through the evaluation criteria used by National Evaluation Bulgarian Longitudinal Inclusive Society and Accreditation Agency (NEAA). By giving Survey (BLISS)37, which is utilizing house- students and other stakeholders the tools to as- hold surveys to assess a broad array of mo- sess the outcomes from institutions and pro- tivation, skills and activation issues across grams, the BURS brings quality issues to the the Bulgarian population. This informa- fore of informed decision-making for both stu- tion could potentially be used as inputs to dents and policy-makers.This is a powerful and funding schemes or in the development of impressive innovation in the higher education other incentive programs. sector for the entire CEE region. The tertiary sector will need to 37BLISS is a nationally-representative multi-topic sur- adapt to an impending decline in en- vey of Bulgarian households, with a booster sample of rollment. As the primary and secondary Roma settlements. It is implemented jointly by the pipeline of students shrinks, so, too, will the World Bank and the Open Society Institute-Sofia and traditional pool of tertiary students. So far, uses the same sample (and includes many of the same modules) as the Crisis Monitoring Survey, which had though the numbers of secondary school stu- three waves (February 2010, September 2010, Febru- dents has decreased, the proportion that goes ary 2011). BLISS has the sample of 2,400 households, to university has grown. With enrollment plus 300 households in the Roma booster. The role of education policies | 27 With increased enrollment of foreign early school leavers to 11 percent; and boost students, classrooms become globalized, the share of persons aged 20–24 years with a bringing in skills and norms from other completed secondary education to 85 per- academic cultures and providing access cent. The first target (5 percent participation to a broader array of modes of teaching in LLL) was missed quite dramatically, as en- and learning. The fees foreign students pay rollments in LLL was only 1.2 percent in could help the financially challenged system. 2012. A very successful life-long learning The students themselves expose Bulgarian stu- strategy has been implemented in Denmark. dents to cultures and ways of thinking that are Key features of this strategy are outlined in different from their home communities. This Box III-1. exposure to foreigners can only help its uni- A Labor Force Survey performed versity graduates to better integrate Bulgaria by the National Statistical Institute in into a global economy. 2011 found that 12.8 percent of young Though Bulgaria has the third-high- adults aged 18 to 24 has left school ear- est median age in the EU, it has the low- ly, close to the goal of reducing the rate est rate of participation in lifelong to 12 percent, but there are notable dif- learning: 1.2 percent. The Bulgarian gov- ferences across groups. The study found ernment recently completed an evaluation of that women left school 1.7 percent earlier the implementation of the National Strategy than men. The worst-performing regions for Lifelong Learning for the years 2008 to are those in the north and south, in which 2013. It provides useful insights on what has 19.4 percent of adults are without secondary worked and what must be improved for the education and 16.8 percent do not partici- next programming period, 2014–2020. The pate in any form of education or training. three leading indicators included in the orig- The Western region, in contrast, maintains inal LLL strategy were: increase participation school-leaver rates between 2 and 3.5 times in lifelong learning to 5 percent of popula- lower than the national 12.8 percent ESL tion aged 25–64 years; reduce the share of rate, with indicators for the period Box III-1: Denmark’s strategy for lifelong learning: A role model for inclusive skill development Denmark has one of the highest levels of participation and completion of lifelong learning endeavors in the EU with 32.3 per- cent of the population aged 25 to 64 participating in training or education program. The reasons for the success of the program lie in: yy A well-developed and properly funded early child education: High-quality basic education has been found to be an impor- tant requirement for successful adult education and training programs through to retirement; yy Publicly funded adult education and continuing training that is inclusive, encompassing general adult education, vocation- al adult education and continuing training (up to and including vocational education and training level--labor market training, basic adult education etc.) as well as a further education system for adults at three tertiary levels. In addition, a number of educational programs are provided for marginalized target groups; yy Public funds provide to all stakeholders - institutions, students, and employers--to support their participation and promote relevant interventions; yy Focused on groups with greatest needs-–the low-skilled, marginalized groups, and those with the lowest levels of formal education and/or literacy and numeracy problems. The aim is that more people with the lowest level of formal education shall participate in vocational adult education and continuing training. Danish Ministry of Education (2007). 28 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria 2008–2011 that are comparable with the learning address the full spectrum of oppor- best values in the EU.38 More challenging, tunities for human capital formation, pro- however, are recent government objectives viding the comprehensive interventions to further reduce the share of early school needed to drive forward effective change the leavers by 2020 to 11 percent. This has been skills make-up of the Bulgarian labor market. established in several subsequent national and European documents including: Na- tional Reform Program (2011–2015), NPD Recent reforms in the education BG2020, National Strategy for Children sector 2008–2018, National Program for Develop- ment of School and Pre-school Education In many countries, populations have be- and Training (2006–2015), Program for De- come older and governments have im- velopment of Education, Science and Youth plemented effective policies for addressing Policies in the Republic of Bulgaria (2009– this change. Many have found education and 2013), Strategy for Educational Integration training to be powerful tools to narrow the so- of Children and Pupils from Ethnic Minor- cial, employment and general-capacity gaps ities of the Ministry of Education, Youth that emerge from demographic shocks such as and Science, National Strategy for Roma those being experienced in Bulgaria. Different Integration in Bulgaria (2012–2020), Law countries have pursued an array of policies to on Youth, as well as within the 28 regional raise the skills of their future work force. strategies for development of social services Table III-3 compares them with what Bulgaria (2011–2015). has done so far. Finally, the National Strategy for The government has taken steps to Life-Long Learning target of increasing help young workers get jobs and to re- the share of persons aged 20–24 with duce the number of drop-outs.The 2010– completed secondary education of 85 2020 National Youth Strategy (NYS) is percent was achieved two years before oriented toward young people aged 15 to 29 the final 2013 evaluation date, when it years and is targeted at “building and imple- scored 85.5 percent in 2011. Overall, for menting a unified, consistent and sustainable the period 2007–2011, Bulgaria experienced youth policy in Bulgaria, based on a multi-sec- four years of improvement and one flat year toral approach, inter-sectoral collaboration followed several years of improvement aver- and joint management with young people at aging 1 percent per year (Eurostat 2012). the national, regional, provincial and munici- The combination of poor learning pal level” (NYS, 2010). The measurable goals outcomes among Bulgarian youth and related to youth and education within the limited participation in lifelong learn- NYS—of reducing the proportion of early ing among adults is a cause for con- school leavers to 11 percent and raising the cern, as noted in the forthcoming World one of university graduates to 36 percent of Bank report of workforce development 39. working age people are comparable with the These combined deficiencies compromise EU 2020 and NDP BG2020 targets. Such pol- the quality of the future workforce and the icy complementarity illustrates the thoughtful country’s ability to move into more lucrative areas of economic activity. At the same time, they diminish the capacity of the current 38 The analysis is based on NUTS3 territori- workforce to adapt to economic restructur- al classification. 39 The World Bank. (forthcoming 2013). Workforce ing and to take advantage of new job oppor- Development in Bulgaria: Systems Approach for Better tunities. Reforms targeting both improved Education Results – Workforce Development (SA- youth education outcomes and lifelong BER-Wf D) Country Report. The role of education policies | 29 Table III-3: Global education responses to demographic shifts Education Intervention Country Bulgaria’s Policy Response Merging/Repurposing Japan, Romania, yyFocal and protected schools identified in primary and secondary school systems of Institutions UK, USA, yyTo be determined at tertiary level Increase Foreign Students Netherlands, yyIncrease the number of programs offered in English and Staff; create Australia, Chile, UK, yyIncrease foreign study options, promote participation in ERASMUS and other incentives for study USA EU education programs abroad and repatriation of skills Lifelong/Life-wide Austria, Germany, yyThe National Concept for Promotion of Active Aging 2012–2030: Learning Hong Kong SAR, yyOperational Objective #4: Ensuring access to education, promoting lifelong China, Ireland, learning, further trainings and requalification of older people for the purpose Japan, Sweden, UK, of enhancing labor market mobility USA yyLaw on Encouragement of Employment—to promote both recognition of life experience in credentialing; promote apprenticeships and skill building in con- junction with employers. Source: Authors’ notes. and collaborative nature of the Bulgarian pol- five-year olds would provide additional icy environment. Likewise, the New Chance opportunities for children of disadvan- for Success project (March 2011) is another taged background to be better prepared program that could bring about important in- for schools and eliminate differences in novations in lifelong educational opportunity. outcome among children based on in- It offers literacy training to early school leavers come. It would also reduce some of the aged 16 and up and to 10,500 working-age known drivers of early school-leaving, such as adults, with the goal of increasing social mo- the inability to cope with the language or the bility. As acknowledged by the Ministry of Ed- curriculum. Pending legislation envisions ucation, Youth and Social Sciences (MEYS), even broader exposure to pre-school educa- “(a)cquiring modern key competencies pro- tion by making it compulsory from age 4. vides a chance for low-literate persons to go Enrollments in pre-schools have already in- out from the risk group on the labor market. creased and will continue to do so if Bulgar- The activities (of the New Chance for Success ia adopts and implements the proposed project) create conditions for inclusion in legislation and reaches for the EU goal of 95 training of persons in disadvantaged socio- percent enrollment into pre-schools of economic status and widen the possibilities for 4-year-olds. Pre-schooling has been proven more successful social inclusion” (MEYS to improve the later scholastic performance 2013). Finally, the government has introduced of linguistic minorities, including the Roma. a Strategy for Prevention and Reduction of To promote lifelong learning, the the Share of Drop-Outs and Early Leavers government developed a LLL program from the Educational System (2013–20) which in 200840. The National Concept for Promo- will monitor drop-outs and early school-leav- tion of Active Aging 2012–2030 highlights as ers. It includes plans to develop a detailed anal- Operational Objective #4: Ensuring access to ysis of gathered data aimed at elaborating education, promoting lifelong learning, fur- long-term policies for the prevention, inter- ther trainings and requalification of older vention, and compensation (remediation of people for the purpose of enhancing labor lost educational opportunities) of early school- leaving. Extending compulsory and free pre- National Strategy for Lifelong Learning for the Peri- 40 school education by one year to include od 2008–2013. 30 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Table III-4: Europe 2020 education goals Goal Examples of BG responses Increase the number of higher The enrollment rate of 30–34years-old was 27.3 percent in 2011 and 26.9 in 2012 according to preliminary education graduates to 40 per- data. cent for 30–34 years old Net enrollments for 19–24-yearolds increased from 26 percent to 42 percent between 2007 and 2011. The net enrollments (bachelor, master and professional bachelor levels) for 20–24 years old increased from 29.5% in 2007 to 35.4% in 2011. Improve the quality and Development of the Bulgaria University Ranking System, providing data on quality and relevance of degree relevance of teacher and programs and institutions. researcher training Provide more opportunities for Foreign student enrollments up 8.1 percent since 2007 (According to NSI data, foreign student enrollments international engagement increased by 21.6% from 2007 to 2011 – incl. all levels of higher education except for the PhD programs. of students and staff—in Foreign student enrollments in PhD programs declined by 12.6% over the same period.) Bulgaria and abroad 22 percent increase in Bulgarian student mobility, 2007–2011* Strengthen the knowledge Bulgaria Smart Specialization Strategy Initiative (ongoing) triangle: teaching, research, business Create effective governance Strengthening Higher Education in Bulgaria: Options for improving the models of governance, quality and funding models in support assurance and financing of higher education (2012 report) developed to provide strategic roadmaps of excellence to achieving effective governance and funding models Reduce early school-leavers yyNational Reform Program (2011–2015); National Program for the development of Bulgaria: Bulgaria 2020; to 10 percent yyNational Strategy for Children 2008–2018; yyNational Youth Strategy (2010–2020) yy28 Regional strategies for development of social services (2011–2015) Source: Authors’ notes based on information from MEYS. market mobility. Beginning in 2014, the next A thorough revision of  Bulgaria’s phase in the lifelong learning agenda41 will teacher policies should lead to  policies call for the Bulgarian National Assembly to that will improve teaching and learning conduct targeted information campaigns to in  the classroom. Currently, 50  percent encourage participation by both institutions of teachers in Bulgaria are over 50 years of age, and individuals in lifelong learning. In partic- the profession is unattractive, and it fails to at- ular, educational institutions are receiving tract the young and the talented. Pre-service support to form consortia to apply for EU- preparation of teachers needs to be improved. funded “Leonardo da Vinci” mobility program Higher education institutes offering degrees funds. The Bulgarian National Assembly has in pedagogy need to revisit their curriculum set a National Priority for 2013 to ensure that and, with Government support, transform a larger number of technical and vocational their programs into highly attractive academ- education and training (TVET) organizations ic majors. Finally, policies ensuring that small participate in the program, so that mobility is schools are attractive to good teachers and en- promoted at all post-secondary institutions, suring that the best teachers are retained not just at elite tertiary institutions. Even if in the system and drawn to schools with sup- they lack the experience and the resources to ply shortages must also be adopted and imple- manage a project at the institutional level, par- mented as soon as possible. ticipating in a broad, well-supported initiative Curricular reforms to promote high- such as this would encourage institutions to er educational standards and matching start building lifelong-learning norms, to offer new curricula and to allow individuals to ac- 41 Being proposed follow-up Strategy for Lifelong cess the new opportunities. Learning 2014–2020. The role of education policies | 31 textbooks are necessary to ensure that Improving educational outcomes students can improve the cognitive and among the Roma non-cognitive skills that employers want. These include critical reasoning, adaptive The Bulgarian government has devel- learning, and the ability to work effectively in oped key education strategies that, if a team. Employers identify them as missing implemented, will prepare its Bulgari- from the current graduates and vital to the an-speaking workforce better for the fu- success of their businesses. Collaboration with ture and mitigate the impact of an aging employers in reforming the curriculum should population. But the expected change in the be encouraged, particularly in TVET, and in country’s ethnic composition also requires a secondary and tertiary education. different kind of reform. As mentioned in the In the context of education and skills previous chapter, between 9 to 19 percent of and in light of Bulgaria’s ageing popula- labor force entrants are projected to be from tion, the challenges for workforce devel- the Roma population. As a result, one of the opment require efforts on four fronts: key priorities is to improve the educational providing better access to education for all outcomes among Bulgaria’s ethnic minori- ethnic groups; ensuring education of high ties. There are very large gaps in educational quality and relevance; boosting worker pro- attainment between the Bulgarian-speaking ductivity through continuing education and population in Bulgaria and the minorities, es- training; and facilitating entry of the inactive pecially the Roma. Among the ethnic Bul- population into the labor force through sec- garian population aged 25 to 64, 59 percent ond chance education and training-related had completed their secondary education, programs.42 Finally, perhaps in a model and 33 percent had post-secondary education like the BURS, pre-primary through (Table III-5). But the figures are 45 percent secondary and TVET education could and 6 percent, respectively, among ethnic benefit from a comprehensive student, Turks (which may also include Roma who teacher and school evaluation system. self-identify first as ethnic Turks), and even Integrating national assessments, classroom lower, 21 percent and 0 percent, respectively, observations, evaluations, inspections and oth- among those who identify themselves er sources, it would provide a powerful source as Roma. of useful information and help guide the de- The Roma’s low educational at- sign and implementation of new policies. tainment is driven by high drop-out Such evidence-based policy-making can pro- mote sustainable progress and help evolving The World Bank. (forthcoming 2013). Workforce 42 systems—such as Bulgaria’s is today—avoid Development in Bulgaria: Systems Approach for Better the pitfalls and wrong turns made in other Education Results – Workforce Development (SA- systems in the past. BER-Wf D) Country Report. Table III-5: Highest level of education attained, by ethnic group in percent Bulgarian Turkish Roma Other Up to basic education - ISCED 1 1 9 34 2 Lower secondary education - ISCED 2 7 40 45 38 Upper secondary education - ISCED 3 59 45 21 42 Post-secondary education - ISCED 4+ 33 6 0 19 Source: 2010 Crisis Monitoring Survey (WB and OSI, 2010); WB calculations. Note: Sample limited to those aged 25–64 years old. 32 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure III-3: Enrolment rates of Roma in Bulgaria by age in 2011 100 92 94 91 Enrollment (Percent) 80 89 87 90 59 60 38 40 40 20 0 10 11 12 13 14 15 16 17 18 Age Source: UNDP/WB/EC regional Roma survey (2011). WB calculations. Bulgaria sample. rates and low pre-school enrollment. Consistent with international evi- Current enrolment data show that around dence, the World Bank (2012d) report 90 percent of Roma in Bulgaria—and in also finds a strong correlation between neighboring countries, with the exception Roma pre-school attendance, gradua- of Romania—remain enrolled as long as ed- tion from secondary schools and atten- ucation is compulsory, but drop out when it dant reduction in drop-out rates. ceases to be at age 16; at age 15, 91 percent Similarly, a representative quantitative study are still enrolled, but only 59 percent at 16 of the Bulgarian International Center for (Figure IV-3). Moreover, pre-school enroll- Minority Studies and Intercultural Rela- ment among Roma children is much lower tions (2003a) demonstrated that Roma than the overall population. Among Roma school children who attended kindergarten children aged 3–6, 38 percent of girls and had significantly higher grades than those 42 percent of boys are reported to be en- who did note. They also were more likely to rolled in pre-school, compared with 75 per- have friends of Bulgarian ethnic origin and cent for the general population. These were twice as likely to be able to use a com- enrolment rates compare favorably with puter. Roma in the Czech Republic (21.8 percent), Finally, the National Strategy for Slovakia (17.8 percent), and Romania the Integration of Roma (2012–2020) (32.2 percent). Only Hungary is a notable aims to increase Roma enrollment, exception, with 66 percent of Roma in this prevent Roma students from dropping age group enrolled in pre-school. out and provide quality education in a Early childhood development should multicultural environment. The Strate- be expanded to improve outcomes for gy also identifies ten specific goals to pro- all ethnic minorities and Roma students mote quality and relevance to the Roma in particular. The World Bank (2012d) re- community. The Strategy aims to integrate gional study on “Closing the Early-Learning Roma students more effectively through Gap” found that Bulgarian Roma children more classroom diversity and the teaching of enrolled in pre-school were much more like- the Roma culture.43 ly to know the Cyrillic Bulgarian alphabet and say simple sentences in the national lan- guage than Roma children in the same neigh- borhood and with similar backgrounds not 43Well-qualified Roma teachers could also serve as enrolled in pre-school. role model for their students. The role of education policies | 33 Policy Options Education policies need to include interventions designed for the Roma Bulgaria’s education policies will have minority. Expanding early childhood pro- to be tailored for today’s and tomor- grams, which have been found to be effec- row’s labor force by expanding LLL tive, will be of particular importance. An opportunities and improving basic and active implementation of the Strategy for tertiary education systems. To enhance Prevention and Reduction in the Share of the productivity of the existing workers, Drop-outs and Early Leavers will be key for Bulgaria will need to invest in LLL pro- reducing high drop-out rates, particularly grams. Identifying barriers for achieving in- among the Roma and other disadvan- creased participation in LLL programs is taged groups. likely to be an important first step. A strong Effective education reforms require collaboration among university graduates, a strong interaction between education training centers and employers would be and labor markets including all rele- helpful for developing strong LLL programs. vant stakeholders by i) identifying the re- To boost the productivity of future genera- quirements of the business community; tions, Bulgaria would also need to improve ii) establishing a tripartite partnership of basic and tertiary education. In particular, university graduates, career centers and em- Bulgaria’s education system should ensure ployers; iii) strengthening the links between that students at all levels learn generic skills higher education and business; iv) selecting that can be used in different work set- the most suitable methods and techniques tings—the so-called transversal skills. In this for recruiting new skilled staff; v) strength- context, improving basic and tertiary educa- ening the coordination between universi- tion through improved teaching and better ties, the non-government sector, and curricula will be important. business. Table VI-6 provides an overview Table III-6: Recommendations for areas of strategic focus and collaboration Responsible Recommendation agency In cooperation/consultation with Timeframe yyIncrease emphasis on generic/ MEYS Association of Rectors, HEIs, Employer organi- Medium-term transversal skills at all education zations, Business associations, Trade Unions, levels National Vocational Education and Training Agency yyEnhance quality of basic and tertiary MEYS National Vocational Education and Training Short/Medium/Long-term education by improving teaching and Agency, Center for assessment of quality, (to learning and through curricular be established) National Education Inspector- reforms ate, Association of Rectors, HEIs, Employer or- ganizations, Business associations, Trade Unions yyIncrease foreign students and staff MEYS Association of Rectors, HEIs, Employer organi- Medium-term in tertiary education zations, Business associations yyExpand lifelong learning opportuni- MEYS MLSP, Employer organizations, business Short/Medium-term ties by identifying current con- associations trade unions straints and fostering coordination among all stakeholders yyUtilize more fully the available data MEYS Statistical agency, Social Security Agency, Short/Medium/Long-term to promote extensive evidence- MLSP based policy making in education Source: Authors’ notes. 34 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria of the different stakeholders involved in the tions and harness education to face the chal- implementation of the different reform op- lenges of its aging population. Indeed, the tions. creativity and purposeful engagement already Bulgarian education is already well- exhibited by the MEYS and the Ministry of supported by an extensive network of le- Labor and Social Policy (MLSP) have set the gal and strategic frameworks helping it stage for an impressive and rapid adaptation. If deal with demographic decline. With the implementation phases of these policy di- continued attention to developing the legal rectives can focus on ensuring good access to and strategic frameworks and roadmaps to im- education for all and on improving the quali- plementation, monitoring, evaluation, and re- ty of existing institutions, then the education calibration (where necessary), there is no sector is well-poised to serve the entire popu- reason Bulgaria cannot build upon its founda- lation of Bulgaria well in the decades to come. The role of education policies | 35 CHAPTER IV Health sector reform in the context of an aging population Health systems and policies play an im- on health, wellness and disease prevention portant role in  shaping demographic programs that target the main causes of mor- transformation. Both can help prevent bidity and premature mortality, in particular unnecessary deaths and ensure that aging obesity and hypertension; ii) coordination of populations enjoy long, productive, and care across health and social services, as well healthy lives. While adults have been living as across different levels of the health care longer and increasingly healthier lives in the system; and iii) recognition of the important EU15  countries, excess mortality, particu- role played by primary care providers in larly in middle age, continues to plague the helping patients optimize their medical care. EU10. Men from the EU10  countries who Many older people are hospitalized for treat- are currently  50  years of  age can expect ments that could be provided more cost-ef- to  live to  age  73–75, compared with fectively as out-patient care. age 78–81 in richer EU countries. Roughly Bulgaria’s rapidly aging population speaking, a  citizen of  Bulgaria is  twice is  expected to  put pressure on  public as likely as a citizen of Spain to die in any health expenditures. The widespread ex- given year44. The health sector not only plays pectation is  that rapidly aging populations a key role in extending lives, it also improves throughout Europe will lead to  increases the quality of  life by  ensuring that people in  health expenditures, simply because the stay healthy and can work. It thus has an im- elderly have a higher demand for ambulato- portant role to play in stemming Bulgaria’s ry, inpatient, and chronic care than younger labor force decline. members of  the population. In  reality, the Preparing a health system for an size and age structure of populations typical- aging population requires concerted ly only explain a  small share of  increases efforts at all levels of the system (Doyle, Rachel, et al. 2009). The complexity of 44 The age-standardized death rate in Spain in 2010 health problems tends to increase as popula- was 488, and in Bulgaria, 970 per 100,000 (1.99 times tions age, with more people suffering from higher). The age-standardized death rate due to car- co-morbidities45 and chronic diseases and diovascular disease in Spain was 138 (28 percent of the total), compared to 622 per 100,000 in Bulgaria thus receiving a wide range of treatments (64 percent of the total). The data are from Eurostat. that could potentially interact with each 45 Comorbities refers to two or more diseases existing other. Adapting to this requires: i) emphasis at the same time in the body. 37 Table IV-1: Health Expenditure - Bulgaria and Comparator Countries (2010) Public Health PHE Health Expenditure Total health expenditure expenditure (percent of gov. (per capita $) (percent of GDP) (percent of GDP) expenditure) Bulgaria 434.9 6.9 3.7 9.8 Romania 428.0 5.6 4.4 10.8 Lithuania 781.4 7.0 5.2 12.6 Latvia 717.6 6.7 4.1 9.2 Poland 917.1 7.5 5.4 11.9 Croatia 1066.7 7.8 6.6 17.7 Estonia 853.3 6.0 4.7 11.7 Source: WHO, Global Health Expenditure Database. Countries in the table are ranked by income per capita. in health expenditure. Whether population trends and suggests priorities for the Gov- aging leads to overall higher health care ernment’s consideration to prepare the costs will depend to a large extent on wheth- health system to meet these challenges. er longer life spans mean more years of healthy living or more years of illness. Fu- ture health care costs are also determined by Current public health expenditure income-driven changes in demand and technological changes in the medical sector. Bulgaria’s total expenditure on health, Bulgaria’s health system is not performing as although low by EU standards, appears well as it could in many areas and Bulgarians to be broadly in line with its income are increasingly dissatisfied with it. This level, but public health expenditures are alone is likely to increase pressure to invest low. Total health expenditure in Bulgaria more in it. It is therefore critical to address represents around 7 percent of GDP, a level some of the system’s structural problems to which is comparable to many EU10 coun- ensure that any further investments will tries. But Bulgaria dedicates a smaller propor- bring about effective results. tion of it public resources to health than does The rapidly aging population adds the average EU country. In 2010, Bulgaria’s a sense of urgency to an already large public health expenditure represented 3.7 need for health service delivery re- percent of GDP and less than 10 percent of forms. Modern health-service delivery sys- total public expenditure (Table IV-1): both tems are under tremendous pressure to adapt figures are rather low by European standards. due to the need to rein in spending, the in- Conversely, out-of-pocket payments (OOP) creasing cost of medical technologies, the incurred by households when they need care evolving expectations of patients, and the represents 44 percent of total expenditures46. imbalances in the workforce. To meet this In fact, Bulgaria is an outlier among EU challenge, systems must be increasingly or- countries: its share of OOP is currently ganized to ensure that patients are treated at among the highest in the region and has the right level of care; there is also a need to concentrate specialized services to provide 46 Out of pocket expenditure is derived from various access to the best quality of care and more sources and pertains to goods and services purchased up-to date complex treatment. This chapter privately: formal public sector copayments and infor- shows that the service delivery system in mal payments, although the latter tend to be un- Bulgaria shows few signs of following these der reported. 38 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure IV-1: Public health expenditure growth and share in public expenditure 2,500 13 12 2,000 11 1,500 10 Percent 9 1,000 8 7 500 6 0 5 2003 2004 2005 2006 2007 2008 2009 2010 2011 Real Health Expenditure (2005 Price) % of Budget on Health Source: Ministry of Finance, IMF. increased substantially over time from around behind most EU countries. Average life 25 percent in 1997. expectancy at birth for a Bulgarian is 73 Following years of decline, public years compared to 80 years in the EU27 and health expenditure growth has acceler- 75 years among the EU10 (Table IV-2). In ated in the past few years. Between 2003 fact, Bulgaria now lags behind most of its and 2009, the share of health in total public neighboring countries, despite performing expenditures eroded steadily (Figure IV-1), relatively better twenty years before. The although it has since increased to its 2003 burden of non-communicable diseases level as public health spending increased 14 (NCDs) is high, in particular circulatory- percent in real terms between 2009 and 2011. system diseases, which tend to mainly affect older adults. In fact, heart attacks, heart fail- ure, and strokes jointly comprise more than Current health-sector performance 65 percent of all reported causes of death in in the context of aging Bulgaria in 2011. Cancer is the next major cause of death (16 percent), with death rates Compared to other EU10 countries, the of 172 per 100,000 similar to the EU-27’s Bulgarian health system’s performance 173 per 100,000 and better than the EU10’s is lagging. A health system’s performance is average 199. measured against 3 goals: i) producing health; ii) providing financial protection47 and iii) re- sponding to citizens’ expectations. Recent ii. Providing financial protection performance reviews of health-care systems suggest that many countries perform better The financial protection provided by than Bulgaria without spending much more. the health care system is limited by high out-of-pocket payments. Financial protection provided by the National Health i. Producing health 47Providing financial protection refers here to protect- Although health outcomes have been ing individuals from the financial consequences of ill- improving, Bulgaria has been falling ness and medical care. Health sector reform in the context of an aging population | 39 Table IV-2: Health Status Indicators of Bulgaria and comparable countriesa SDR, Standard diseases of Life Reduction of Death Rate the SDR, SDR, chronic expectancy LE through Infant deaths (SDR) all circulatory malignant liver disease/ Tuberculosis (LE) at birth, death before per 1000 live causes, per system, per neoplasms, cirrhosis, per incidence per Countries in years 65b births 100000 100000 per 100000 100000 100000 Austria 80.6 4.1 3.8 563 213 158 15 5.3 Bulgaria 73.4 6.9 8.6 995 611 172 18 35.4 Czech 77.5 4.8 2.9 744 357 197 16 6.0 Republic Greece 80.3 4.0 3.2 577 245 154 5 4.1 Romania 73.6 7.2 10.1 959 549 181 47 97.2 Slovakia 74.3 6.3 7.2 945 509 208 25 8.1 Slovenia 79.3 4.6 2.6 632 235 202 25 9.1 EU 79.6 4.6 4.3 622 234 173 14 13.5 EU10 75.1 6.6 6 873 436 199 27 35.9 Source: HFA database (accessed Aug 2012). a Hypothetical increase in life expectancy if no one died before 65. b Data is for 2009 or latest available year. Insurance Fund has significant gaps and be- medicines as a crisis-coping mechanism; the tween 10 and 20 percent of the population average for all households was 19 percent. are not covered.48 As a result, out-of-pocket Overall, this shows that out of pocket pay- payments (OOPs) in Bulgaria are large. In ments, which are predominantly incurred 200749, they represented nearly 6 percent of on pharmaceuticals, are more detrimental to average household spending. Two thirds the poor. were spent on medicines, which are not well The elderly are thus particularly covered by the public health insurance sys- vulnerable. The elderly are currently at the tem. In the health economics literature, highest risk of poverty and social exclusion OOPs in excess of 10 percent of household in Bulgaria51. They are also more likely to spending are deemed catastrophic. In Bul- develop chronic diseases than younger age garia, 20 percent of households faced OOPs groups. Controlling for risk factors associat- above this threshold compared to only ed with chronic diseases or managing these 7 percent of households in EU15 countries. diseases typically requires taking medicines The same survey showed that OOPs have on a long-term basis. For these two reasons, had a significant impact on poverty in Bul- the burden of out-of-pocket payments prob- garia: using a poverty line of $5 per day, ably falls disproportionately on the elderly. OOPs on health were responsible for in- As their number increases, the need for im- creasing poverty from 12 to 15.7 percent50. proved financial protection can be expected A three-wave household survey conducted to rise. in 2010 and 2011 showed that many house- holds reduced their spending on health in 48 No official data available. response to the economic crisis (World Bank 49 Last Living Standards Measurement Study available. 50 World Bank computations. 2012a). In fact, 40 percent of those house- 51 2011 SILC data http://epp.eurostat.ec.europa.eu/ holds in the lowest-income quintile declared c ache/I T Y_ PU BLIC/3 –26 02 2 013 -AP/ EN/3 – having suspended the purchase of regular 26022013-AP-EN.PDF. 40 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria iii. Responding to citizens needs 2009). The Euro Health Consumer Index53 (Health Consumer Powerhouse 2012) ap- A significant majority of Bulgarians has pears to confirm Bulgaria’s poor user-focus: it expressed dissatisfaction with the health scored 33rd out of 34 countries in Europe, a care system. In 2009, only 28 percent rated virtual tie for worst with Serbia. the system as “good” or higher: the second- lowest rating in the EU (European Commis- sion 2010). One possible reason for A service-delivery agenda with a dissatisfaction is unmet needs: In the 2008 focus on the hospital sector Statistics of Income and Living conditions survey, almost a quarter of Bulgarian respon- i. Improving coverage of preventive dents reported an unmet need for medical services, use of primary care and consultation, with the most commonly cited hospitalization rates reason being cost. By 2010, the rate had dropped to around 15 percent for an EU av- Poor health does not have to be an in- erage of around 7 percent52. Perceived cor- evitable consequence of aging: healthy ruption within the health care system could behavior and preventive health care be another reason for such low levels of satis- can help people live long, healthy lives. faction: a 2009 survey found that 65 percent In Bulgaria, the of most preventive services of Bulgarians believed corruption was wide- is much lower than in other EU countries, spread in the public health care sector—slight- with the exception of Romania (Table IV-3). ly higher than the EU12 average of 54 percent, but virtually the same as the corrup- 52 Hungary 7.9 percent, Slovakia 5.5 percent, Romania tion perceived in other public sector domains: 13.6 percent (Statistics of Income and Living Condi- tions survey, via EUROstat). building inspections, public contract tenders, 53 A compound measure of nearly 40 indicators which licensing inspections and business permits, all assesses patient rights, health services and pharmaceu- at 60–65 percent, (European Commission tical accessibility, as well as health outcomes. Table IV-3: Use of preventive services in Bulgaria and selected EU countriesa Breast exam Colorectal cancer Cervical smear Cervical Cancer Influenza (women 50–69) screening age test (women death rateb immunization (percent) 50–74 (percent) 20–69) (percent) (per 100,000) (percent) Bulgaria 10.3 8.5 18 7.9 4.8 Czech Republic 39.8 14.1 46.3 4.9 19.4 Germany 44.7 36.6 58.3 2.5 56.2 France 50.3 13.6 48.7 1.9 66.7 Latvia 23.1 8.5 41.7 5.9 2.9 Hungary 37.4 2.9 35.4 5.7 30.3 Poland 29.4 1.7 35.8 7.1 12.9 Romania 3.5 0.7 4.4 13.4 18.1 Slovenia 25.6 3.3 38.5 3.7 22.3 Slovakia 31.9 9.6 33.1 6.3 24.4 Source: European health interview survey, most recent year available (circa 2008). a Percentage reporting use of preventive services within the past 12 months in Bulgaria and selected EU countries. b Eurostat 2009 or 2010. Health sector reform in the context of an aging population | 41 One example is women’s health: only 10 a plan to reinforce the prevention of non- percent of women aged 50–69 received communicable diseases. mammography screenings in Bulgaria com- But the system is not addressing pared to 23 percent in Latvia and 37 percent the burden of disease in an efficient in Hungary. Only 18 percent of women way. Though Bulgaria has a high number of aged 20–69 had a cervical smear test in Bul- physicians per capita compared with other garia, compared to about 35 percent in East- countries, Bulgarians have fewer contacts ern Europe and about 50 percent in Western with primary care and specialist physicians Europe. Death rates from cervical cancer re- than citizens in other EU countries with the main high in Bulgaria but could be practi- exception of Romania. The number of cally eliminated by early detection and nurses per capita is by far the lowest in the treatment. Influenza vaccine coverage in region. There are 4 nurses per thousand in- Bulgaria is another illustration of the limited habitants in Bulgaria, 6 on average in EU10 focus on prevention care. The World Health countries and 9 on average in EU15 coun- Organization recommends immunization tries. Many nurses have emigrated and in for some categories of the population, in- 2011, more physicians than nurses com- cluding the elderly. In Bulgaria, coverage is pleted their studies. Further, the number of less than 5 percent, compared to about 20 hospital episodes per capita jumped sharply, percent in Eastern Europe and more than 50 by 68 percent between 2000 and 2010, as percent in Western Europe. other countries decreased hospitalizations The Ministry of Health has stepped (Table IV-4). As mentioned in the introduc- up efforts to improve prevention care in tion, for reasons of efficiency and quality, recent years. Most notably, in 2010, the Hib most countries aspire to reducing hospital- (Haemophilus influenzae type B) and the izations by relying on a combination of PcV (Pneumococcal conjugate) vaccines prevention and outpatient treatment, par- were introduced in line with recommenda- ticularly for chronic disease management. tions from the Centers for Disease Control Bulgaria appears to be moving in the op- and Prevention and other international agen- posite direction. Bulgaria’s hospitals are pri- cies. In 2010, tobacco tax rates were increased marily funded on a case-basis (meaning they sharply, and bans on public smoking were in- receive a payment for each stay), which stituted in 2012.These are two commendable provides strong incentives to multiply hos- policy actions that have had positive proven pitalizations. Expenditure caps have been public health impacts in other countries. The put in place to counter-balance this but Ministry will launch a National Program for strong incentives remain for hospitals to Cancer Screening in 2013 and it is finalizing encourage hospitalizations. Table IV-4: Inpatient hospital discharges Bulgaria and select EU countriesa Countries 1990 2000 2005 2010 Bulgaria 19.0 15.4 21.0 25.9 Croatia 15.4 15.7 16.6 16.8 Hungary 21.8 23.6 25.0 20.8 Romania 20.1 22.4 24.6 24.9 EU 15 16.9 17.7 16.9 16.9 EU10 16.8 19.0 20.8 21.0 Source: HFA database. a Per 100 persons. 42 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure IV-2: Acute care hospital beds per 100thds 700 Lithuania 650 Latvia 600 Czech Republic Hungary 550 Slovakia Bulgaria 500 Slovenia Poland 450 Estonia 400 Croatia EU15 350 300 United Kingdom 250 200 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Bulgaria Czech Republic Estonia Hungary Latvia Lithuania Poland Slovakia Slovenia United Kingdom EU15 Croatia Source: Health for All Database 2011. ii)  Reducing size and fragmentation Figure IV-3: Typology of hospitals of the hospital system in Bulgaria Among EU10 countries, Bulgaria has the Other or highest number of beds per capita and, Unknown 6% opposite to the trend in most other countries, the number continues to grow. Despite large cuts in hospital infrastructure in Long-Term Care Facilities the 1990s, EU10 countries and Bulgaria in 12% Multi-Profile Acute particular, still have a relatively high stock of Care Hospitals 52% hospitals and beds. Figure IV-2 shows the trend in the number of beds per 100,000 since Specialized Hospital 2000, with the United Kingdom at the bot- 30% tom for reference. The EU10 and the EU15 are all positioned above the UK. A significant share of Bulgaria’s hos- pitals is too specialized to provide the multi-disciplinary quality required for an aging population. The large share of Source: Author’s calculations based on National Center of specialized hospitals is a legacy of the Soviet Health Informatics Data. system on the one hand, which relied heavily on the delivery of services through single- specialty hospitals. On the other hand, many of the 275 hospitals contracted by the NHIF new hospital entrants are specialized institu- are specialized. tions opting to focus on delivering the most Many Bulgarian hospitals have a lucrative services. Currently nearly one third very low throughput, which is highly Health sector reform in the context of an aging population | 43 Figure IV-4: Monthly distribution of number of Bulgarian hospital patients 200 In a country where service delivery consolidation is advanced, small acute care hospital 150 might typically treat at least Number of Patients per Day 13 patients a day In Bulgaria, 103 hospitals discharge 100 less than 5 patients a day 50 In Bulgaria, 192 hospitals discharge less than 13 patients a day 0 #136 #123 #193 #17 #197 #244 #106 #34 #261 #84 #69 #47 #60 #159 #1 #99 #198 #92 #139 #238 #110 #144 #233 #124 #243 #177 #224 #90 #10 #95 #108 #44 #102 #153 #6 #256 #113 #112 #225 #201 #66 #59 #160 #105 #191 #158 #94 #93 #214 #152 #192 #275 #172 #223 #135 #42 #28 #179 #200 #241 #37 #240 #101 #36 #3 #264 #104 #140 #96 #12 #67 #236 #168 #119 #182 #181 #235 #46 #109 #63 #85 #170 #245 #41 #19 #255 #21 #9 #216 #97 #232 #267 #31 #204 #8 #263 #82 #132 #116 #213 #120 #269 #89 #30 #254 #111 #39 #76 #167 #57 #165 #24 #222 #11 #27 #217 #257 #156 #242 #231 #22 #162 #180 #50 #211 #239 #206 #14 #79 #71 #166 #13 #40 #122 #80 #52 #163 #16 Facilities Ranked By Number of Patients Source: Author’s Calculation based on National Center of Health Informatics Data. inefficient. Figure IV-4 shows the 275 fa- coordination of investment and thus the ca- cilities ranked by number of patients dis- pacity to finance and concentrate investment charged per month. The three busiest in technology-intensive medical equipment. hospitals discharge more than 95 patients a Technology is one of the keys to improving day. At the other end of the spectrum, the the performance of the health-care system first 103 facilities account for 5 percent of (Smith, Mossialos, Leatherman and Papanico- hospital stays in Bulgaria, which means that los 2008). In one study, information was col- approximately one in three hospitals in Bul- lected about medical equipment available in a garia discharges 5 patients a day at most. sample of 10 hospitals. It revealed that that a Further, if the smallest acute-care hospital in supposedly sophisticated multi-profile hospi- Bulgaria had 130 beds54, it could be expect- tal had less equipment than a lower-level fa- ed to discharge around 13 patients a day. In cility and a so-called “Comprehensive Bulgaria, 192 (out of a total of 275) discharge Oncological Center” appeared to lack the fewer patients than that55. The hospital sys- tem is therefore fragmented, in other words, 54 The assumptions underlying this simulation are composed of an exceedingly large number plausible. For instance, in the Netherlands, in 2011, the of facilities that produce very little (and are smallest hospital has 138 beds (the threshold used here probably small). There is considerable room is 130). Historically, this was not the case: In the 50s, more than half of Dutch hospitals had less than 130 for consolidation in the sector. beds, but the system has since been profoundly remod- The fragmentation of hospitals eled. Other assumptions are as follows: the occupancy tends to lead to the duplication of re- rate of beds is 70 percent (which is rather low – in 2008 sources across facilities, impeding the in Bulgaria it was close to 76 percent) and each patient exploitation of economies of scope and stays 7 days in the hospital (the average length of stay in Bulgaria in 2011 was 5.8 days). scale that modern health care requires. 55 Even accounting for long term care hospitals (less Competition between facilities hinders the than 35 out of 275) where stays would be longer than coordination of care; it also hinders the in acute care hospitals. 44 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria technology required for this type of facility56. Referral patterns highlight some The 2012 Health Systems in Transition re- dysfunctions and suggest that the qual- view (Dimova et al. 2012) reports that the ity and effectiveness of care might be material and technical infrastructure of state sub-optimal. In theory, an efficient health and municipal health establishments in Bul- system should be organized as a pyramid garia is obsolete and that substantial funding through which patients travel and are elevat- will be required to modernize buildings and ed when the complexity of their problem re- medical equipment.57 The consolidation of quires more specialized resources. On paper, facilities and investments in medical equip- this is the case in Bulgaria but in reality, ment will need to be strongly coordinated. there are clear signs that it is not happening. For instance, nearly a third of cases are di- rectly referred from general practitioners, iii)  Improving the allocation of which is much higher than one would ex- patients to the right setting pect. Another study (Sanigest Europe 2008) found that only about half of the total ad- Around 20 percent of hospital stays in missions were planned and the other half Bulgaria are for conditions which, in- (49.3 percent) were urgent (whereas in the ternational standards suggest, could be UK, for example, about 80 percent of ad- routinely treated on an outpatient ba- missions are planned). A number of observa- sis.58 The list of 308 Bulgarian Clinical Care tions also suggest that the hospital system Pathways (CCPs—see footnote 13) was re- may not be treating patients effectively. First, viewed by an international team of hospital at least a quarter of patients are “referred physicians who estimated that 49 CCPs cor- from other hospitals”. If this is genuinely the respond to cases usually amenable to out-pa- tient care, according to international good 56 A physical assessment of the state of medical equip- practices59. Between 60 and 100 percent of ment and an audit of their actual use would be required these episodes could be dealt with without for a complete picture of the state of technology in these hospitals. an overnight stay in a well-functioning 57 Another important drawback of specialized hospitals health care system with properly-staffed pri- is that they are often not financially sustainable and, at mary health care facilities, effective emer- the level of individual public facilities, the accumula- gency care and a substantial network of tion of arrears and debts has become a recurring issue. “community care” providers, as well as a Data on hospital debts and arrears are not readi- ly available. balanced payment system which provides 58 Data analyzed here are from the records of each hos- incentives to treat patients at the right level. pital stay reported by hospitals to the National Health Given these conditions, the data suggests insurance Fund (NHIF) and collected by the National that that one in five hospital stays could be Centre for Public Health & Analysis between July 2011 and June 2012, a total of 1,558,339 anonymous avoided in Bulgaria. This proportion of cas- records. These records come from 275 acute and long es corresponding to CCPs amenable to out- term care hospitals. Virtually all hospitals with a con- patient care varies significantly across tract with the NHIF are included with the exception hospitals. Approximately 40 percent of hos- of the military hospitals and mental health facilities. pitals do not appear to provide these CCPs. Each record includes information about the patient’s demographic characteristics; diagnoses; medical pro- At the other end of the spectrum, in 20 per- cedures performed; length of stay. It also includes the cent of hospitals, more than a third (and up Clinical Care Pathway (CCP) selected for the stay. The to 100 percent) of the activity could be ame- CCP is the flat-rate payment which is paid to the hos- nable to outpatient care. Interestingly, there pital for a complete stay, there are 308 different CCPs. 59 These correspond to chronic conditions which could is no correlation between a hospital’s size be treated on an outpatient basis, conditions for which and the proportion of activities that could be outpatient surgery is a reasonable option and condi- carried out on an outpatient basis. tions related to rehabilitation or palliative care. Health sector reform in the context of an aging population | 45 case (and if the treatment indeed required a regulations which may not encourage or hospital stay in the first place), it means that even allow outpatient alternatives; and the the first facility failed to resolve the prob- lack of incentives to change what has be- lem, which should have been the case in the come normal practice. Second, the propor- vast majority of cases. The data base also tion of cases amenable to ambulatory care suggests that around 13 percent of the popu- does not appear to be correlated with hospi- lation had at least one hospital stay in the tal stated complexity or size (measured by year. For these patients, the average number activity) which suggests that hospitals are of hospital stays is 1.6, and around 30 percent not organized in networks and thus all com- of patients were hospitalized more than once pete with each other to provide all services in a given year60. In the end, the fact that rather than try to complement one another. many patients are hospitalized more than once in a year is probably a sign that (i) some of the treatment did not warrant a hospital Policy Options stay; and (ii) the treatment provided may not be adequately effective. The introduction of the health insurance The continued inefficient allocation fund in 2000 in Bulgaria was seen as a of patients is likely to lead to a signifi- decisive move away from the legacy of cant waste of resources. Table IV-5 at- the communist regime and towards a tempts to quantify the impact of the stays West European system (Atanasova et al. amenable to outpatient care (as defined 2011). Expectations have not been met and above) in 5 representative multi-profile acute Bulgarians are increasingly dissatisfied with hospitals. The fact that the average length of their health care. Resource constraints could stay for these cases is always well above 4 days possibly explain the lack of improvement, but illustrates the extent to which the system is countries with comparable economic circum- biased toward in-patient care. In these five stances appear to have achieved better results. hospitals alone, a significant proportion of The imbalance and ineffective organization of around 200,000 bed/days could be avoided. the Bulgarian service-delivery sector is at the Together, these findings highlight the many ways in which health-service 60Data on the number of times given individuals are delivery needs to change. First, a signifi- hospitalized in a year is not readily available in most cant number of patients are being admitted countries. In France in 2008 around 13 percent of the for problems which could be taken care of residents were hospitalized once or more (a rate com- parable to that of that Bulgaria). Yet, total number of on an outpatient basis. The root causes are hospitalizations per capita was 40 percent lower in likely to include the limited availability of France, meaning that the number of patients with outpatient alternatives; payment systems and multiple hospital stays is unusually high in Bulgaria. Table IV-5: Activity amenable to outpatient care in 5 multi-profile hospitals Number of Percent of activity Hospital Level Average LOS Min LOS Max LOS days for the concerned hospital A 2nd 4.0 1 14 10,091 36.9 percent B 3 rd 4.8 1 106 60,749 24.4 percent C 1 st 5.2 3 16 2,076 16.9 percent D 3 rd 4.8 1 81 68,114 17.9 percent E 3 rd 5.1 1 89 50,895 30.1 percent LOS: length of stay. 46 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria core of the issue, and as illustrated earlier, in due to health care costs—is a core ob- the absence of reform, the problems will only jective of the health system. As men- get worse as the population ages. There are tioned above, financial protection provided no easy or quick solutions. Transforming by the National Health Insurance Fund has the health sector will require decisive ac- significant gaps and OOP expenditures are tion on several fronts: large in Bulgaria. It is important to note that A key action would be the imple- both large OOP expenditures on drugs, mentation of a technically-driven hospi- which are more problematic for the poor tal rationalization plan to support the and the elderly, and exclusions from the reg- reorganization of the hospital sector ular insurance system are sources of ineffi- while maintaining physical access to ciency in the system. Indeed, patients who care. In this context, the National Health In- forego care or do not manage their existing surance Fund needs to be able to purchase conditions are at higher risk of complication services selectively, i.e., it should be allowed to and ultimately end up costing the public sys- decide which entities it can contract. To sup- tem more. In addition to being equitable, port this process, information on the quality improving financial protection is efficient. of care needs to be generated, collected and Although the number of physicians publicized. As shown above, the CCPs gener- per capita is high, the profession’s spe- ate a host of poor incentives. The implemen- cialty mix is not adapted to the popu- tation of a modern Diagnostic-Related Group lation’s needs. The number of nurses per (DRG) payment system has been under con- capita is by far the lowest in the region. So- sideration for many years, but efforts to imple- lutions could include interventions on vari- ment this have not prevailed. Accompanying ous fronts, including planning for human the introduction of DRGs with expenditure resources, adapting training, and addressing caps and the development of outpatient alter- financial and other constraints in order to natives through concerted efforts (regulatory retain qualified staff in Bulgaria. and in terms of payment systems) would avoid Addressing the various issues high- further escalation of costs. lighted in this chapter will require Strengthening alternatives to hospi- strong and consistent strategic leader- tal-based care would require improving ship. Transparent debates that articulate the capacity of primary care health pro- clear priorities are likely to improve trust in fessionals to manage the prevailing bur- the health system over time. A forthcoming den of disease and to play an active role report on Health Systems in ECA highlights in the coordination of their patients’ that while there are no specific recipes to care needs. This would require developing building better health systems, the use of an effective and attractive continuous medical information for decision making and strong education curriculum in light of the current leadership are crucial. As the report puts it: burden of disease and adjusting regulatory “Successful reform requires vision and lead- standards to expand the list of conditions that ership. It means taking on vested interests, can be managed fully in primary care. Stron- whether in the medical establishment, po- ger incentives and accountability mechanisms litical actors, or elsewhere in society, to would also need to be put in place to increase usher in new reforms that will help achieve the management of chronic diseases at the sector objectives.” Many experts within the primary care level (pay for performance). health sector in Bulgaria agree that the sys- Emergency care reforms could help improve tem is in need of reforms and support the the continuity of and access to care. recommendations presented here, but at- Providing financial protection to tempts at reform have so far not been all—so that no one falls into poverty brought to completion. Health sector reform in the context of an aging population | 47 CHAPTER V Long-Term Care Expenditures and Policy Options One of the most important alternatives detrimental to the well-being of older people, to hospital-based care in the context of because they are exposed to infections in hos- population aging is the development of pitals. A better solution would be to discharge the long-term care (LTC) sector. The the patient when no more medical treatment relatively high average length of hospital stays, is needed and to provide the social support discussed in the previous chapter, suggests through community-based LTC services. that improved patient discharge management There is a strong consensus in the could cut hospital costs. The aim of patient literature that population aging signifi- discharge management is to reduce the length cantly increases the demand for long- of stay in hospital by referring patients to term care (LTC). The steep decline of the more cost-efficient LTC services, either with- working-age population in Bulgaria raises in the health system (outpatient services like the question: who will care of the elderly post-surgical treatments and rehabilitation) or and who will pay for future LTC needs? The in the social system. The fact that Bulgaria urgency of policy action stems from the currently spends only 1.2 percent of total need to plan future financing needs for LTC health expenditures on rehabilitation suggests services now. While today’s workers may be a lack of rehabilitative services. In contrast, in able to finance LTC needs for their parents’ other EU10 countries, where such data are generation through a pay-as-you-go system, available, expenditures on rehabilitative ser- this is unlikely to be sustainable for future vices as a share of total health expenditures is generations. Given current demographic around 2.9 percent. Hospital discharge man- trends, today’s workers will not be able to agement of older patients could also involve rely on their children to finance their LTC social workers who can help organize assis- needs. Bulgaria, therefore, needs to find tance with ADLs after discharge. For older ways to increase earmarked savings for the patients who live on their own and lack fam- future LTC needs of today’s workers. Finan- ily support, even a minor injury like a broken cial LTC insurance, like that implemented leg can leave them temporarily dependent on in France, is one option that could be ex- outside help. Given the overcapacities in hos- plored to achieve this. pital beds, hospital management may choose People in need of LTC services are to keep such patients hospitalized until they not only elderly patients but anyone are fully recovered. This is expensive and also with restrictions in performing activities 49 of daily living (ADLs). They are hampered tion in nursing homes); personal care in their mobility and in performing such ac- services to assist with ADLs; and services tivities as cooking, cleaning, washing, shop- and financing in support of informal (fami- ping and eating.They are therefore dependent ly) care; and (ii) in the social sector: home on others to support them in these activities. help and care assistance; residential care ser- Their dependency could be the result of a vices other than nursing homes; and other mental or physical disability, but in the case of services like daycare and transportation. the elderly it is not necessarily disability, but Available data suggests that Bulgar- frailty in general that places them in a position ia and other EU10 countries spend rel- of dependency. In Bulgaria, only 0.7 percent atively little on LTC. Total public and of the population aged 16 to 24 is considered private expenditure on LTC in Bulgaria is highly dependent, while the same share is shown as ranging between nil and 0.2 per- 14.5 percent for the population aged 75 to 84, cent of GDP,61 placing it among the Europe- and 23.9 percent for the population aged 85 an countries that spend least on LTC. The and older (Table V-1). These shares are some- data on LTC available in the OECD SHA what lower than in other EU countries. Still, database shows considerable variation across it is important to constantly assess population countries and time. High-income countries growth rates among the old and very old, with like Sweden currently spend as much as 3.7 an aim to ensuring continued quality of care, percent of GDP on LTC. In contrast, the since they place most demand on LTC EU10 countries report spending less than 1 services. percent of GDP on long-term care. The data also shows strong variations over time for some countries. To illustrate, Latvia record- Current public expenditures on ed expenditures on LTC of LAT111.04 mil- long-term care lion in the social sector in 2005, while in 2006 this expenditure category is missing. In general, there are few data available Countries probably spend more on on current public expenditures on LTC than the SHA data suggests. The long-term care. This makes it particularly SHA data on LTC has a strong bias towards difficult to compare expenditures across health sector data, with data on expenditures countries. The undefined position of LTC in the social sector largely absent, especially between the health and the social sectors for EU10 countries. Focusing on public LTC also makes it difficult to accurately collect expenditure can help overcome this bias data. The OECD (2008) developed a System though at the expense of narrowing the scope of Health Accounts (SHA) where LTC com- of the analysis. In most countries, the largest prises the following categories of services: (i) in the health sector: palliative care; long- term nursing care (including accommoda- 61 Excluding the outlier in 2005. Table V-1: Share of highly dependent people by age group in 2008 16–24 25–34 35–44 45–54 55–64 65–74 75–84 85+ Bulgaria (2011) 0.7 1.3 1.3 1.6 5.1 7.3 14.5 23.9 Bulgaria (2008) 1.0 0.9 1.6 3.1 5.9 10.1 14.9 30.1 EU-average (2008) 1.5 2.3 3.8 6.9 10.9 14.6 25.0 39.5 Source: Eurostat (2013). Note: Based on self-reporting. 50 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria share of public expenditures for LTC in the 0.4 percent estimated by the SHA. Table V-2 social sector, in particular for the elderly, oc- lists total expenditures on LTC (as a percent- curs at the level of local government. This de- age of GDP) only for countries where expen- centralized financing of LTC services makes it ditures are relatively constant over time, which difficult to obtain consolidated national ex- suggests at least consistency of data collection penditure data. A recent initiative by the within countries. World Bank focuses on developing databases Bulgaria’s public expenditure on for consolidated government expenditures to LTC in social sectors is estimated to include expenditures at all government levels. amount to 0.22 percent of GDP in 2012. This will enable the collection of detailed Total public expenditures on LTC increased public expenditure data on LTC at the local somewhat from BGN 154 million in 2010 to level as well as consolidated administrative BGN 168 million in 2012 (Table V-3). Total data from the Ministry of Social Policy. A first public LTC expenditures include expendi- detailed analysis of such data was recently tures for incapacity, temporary incapacity, conducted in Poland. It showed that estimat- disability, and support for the elderly in the ed public expenditures on LTC in the health social sector, but exclude any spending on and social sector were at 1.0 percent of GDP LTC in the health sector or any private in 2007, in contrast with estimated total ex- spending on LTC. Cash benefits or social se- penditures, including private expenditures, of curity benefits, unless specifically assigned Table V-2: Total LTC expenditures as a share of GDP in selected countries Average 2003 2004 2005 2006 2007 2003–2007 Sweden 3.8 3.8 3.7 3.7 3.6 3.7 Germany 2 2 2 1.9 1.9 1.9 Japan 1.6 1.7 1.7 1.7 France 1.4 1.5 1.5 1.6 1.6 1.5 Austria 1.3 1.3 1.3 1.3 Slovenia 1.1 1.1 1.2 1.2 1.1 1.1 United States 1 1 1 1 1 Latvia 1.4 0.2 0.8 Spain 0.7 0.7 0.7 0.7 0.8 0.7 Lithuania 0.4 0.4 0.4 0.5 0.4 Poland 0.4 0.4 0.4 0.4 0.4 Hungary 0.3 0.3 0.2 0.2 0.3 Czech Republic 0.3 0.2 0.2 0.3 0.3 0.3 Estonia 0.1 0.1 0.2 0.2 0.2 0.2 Portugal 0.1 0.1 0.1 0.1 0.1 Bulgaria 0 0.1 0.2 0 0 0.1 Romania 0 0 0 0 0 0 Average 1.2 1.2 1.2 1.1 1.1 1.2 Source: Eurostat. Note: The 2005 value for Hungary (8.23 percent of GDP) was dismissed as an outlier. Long-Term Care Expenditures and Policy Options | 51 Table V-3: Government expenditure on LTC according to government expenditure data (BGN) Capacity Expenditures Expenditures per capacity 2010 2011 2012 2010 2011 2012 2010 2011 2012 LTC institutional Homes for elderly 5,318 5,427 5,462 30,014,946 31,621,851 31,036,724 5,644 5,827 5,682 Homes for adults with 4,159 3,734 5,317 40,346,110 39,072,828 37,765,306 9,701 10,464 7,103 disabilities Centers for social 1,958 2,547 3,139 4,754,425 5,833,034 6,687,367 2,428 2,290 2,130 rehabilitation and integration LTC community-based Domestic social 38,030 38,390 38,554 43,228,311 45,540,252 46,569,962 1,137 1,186 1,208 patronage, public canteens Clubs for pensioners, n.a. n.a. n.a. 6,283,563 7,476,566 6,584,993 n.a. n.a. n.a. disabled people Daycare for elderly 1,388 1,413 1,423 2,493,591 2,760,346 3,030,603 1,797 1,954 2,130 Daycare for adults and 2,376 1,913 3,388 15,877,149 17,799,745 18,706,607 6,682 9,305 5,521 children LTC home-based Foster care for children 627 993 1,286 5,897,778 8,889,981 11,377,950 9,406 8,953 8,848 and adults Shelters 875 959 941 5,415,328 5,975,322 6,488,754 6,189 6,231 6,896 Total 54,731 55,376 59,510 154,311,201 164,969,925 168,248,266 2,819 2,979 2,827 Source: Administrative data. to support care or care givers, and most im- International experience shows that portantly old-age or disability pensions, are many countries at some point in their also excluded. history have converted redundant mu- nicipal hospitals into LTC institutions. These reforms however bear that risk of in- Providing long-term care services troducing a bias towards expensive institu- tional care into a country’s LTC system. Reforms that focus on community- Poland is a good recent example. Some years based services like home care and day ago, Poland started to convert small hospi- care for the elderly are likely to lead to tals into medical nursing homes, run and fi- more efficient and produce better out- nanced by the health sector, that were comes. Institutional care is an important intended to provide post-surgical treatment component of any LTC system, but it results at lower costs than in regular hospitals. The in higher-intensity care and is more expen- unintended consequences were that, given sive; it is also often not the preferred form of the general shortage of LTC services in Po- care by patients. The conversion of small land, patients and their families continued to municipal hospitals into institutional LTC use the medical sector as a substitute for so- facilities is a good example of how well-in- cial LTC services, at a much higher price tentioned LTC reform can lead to ineffi- than this could be done in the social sector. cient outcomes. This came about because medical nursing 52 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria homes were largely financed by the health Financing of LTC services insurance fund and came at a much lower price for patients than private or social sector Some form of pooling of LTC risks is LTC services. Private sector response was desirable from an economic perspec- dampened because for-profit and even non- tive. As demand for LTC services is expect- profit organizations could not compete with ed to increase dramatically, so are its costs. the lower user fees in medical nursing homes. Individuals who lose their ability to live in- Municipalities, who finance most of the so- dependently are therefore at risk of becom- cial LTC services, also found it cheaper to ing impoverished. Evidence from Austria shift patients to the health sector than to suggests that the estimated income and assets provide their own social services. of an average Austrian pensioner would fi- A better approach would be to con- nance only about four months of institution- vert redundant municipal hospitals into al LTC. This compares to an average length privately or publicly owned community of stay of 35 months. As the need for LTC is centers that provide a whole range of a low-probability, high-cost event, econom- LTC and rehabilitative services. Such ic efficiency requires some form of risk- community centers would be at the center of pooling to allow people to protect themselves care coordination for patients. They could against the adverse financial impact of LTC house day-care centers for elderly and disabled needs. Hence, not only for reasons of equity, people (or even childcare), but also offer out- but also for reasons of economic efficiency, patient services like physical therapy. They individuals should pool their resources to could host home-based services like care assis- protect themselves against the risk of be- tants or community nurses who support de- coming dependent, just as they should pro- pendent people in their homes. To the extent tect themselves against the risks of sickness.62 necessary, they could of course also provide For various reasons related to asymmetric limited facilities for residential care, in particu- information, risk selection, adverse selec- lar respite care and temporary care. The exact tion, and cost uncertainty, the potential for setup of such community centers would de- private LTC insurance to provide in-kind ser- pend on local needs and circumstances and vices is limited and in practice, it has not also have to be balanced against transportation been very successful.63 This suggests that costs in sparsely populated areas. High trans- there is a need for the public sector to play a portation costs could in fact justify more resi- role in providing adequate instruments for dential care facilities with low levels of pooling old-age dependency risks. care intensity. The most common forms of pub- Finally, provider payment mecha- licly provided risk-pooling are financed nisms should ensure a level playing by contributions to social security and field for competing public and private health benefits, and tax-financed bene- providers. If public providers are subsi- fits like social safety nets. For LTC, most dized, either directly from the municipal countries use both mechanisms to finance budget through staff salaries or through cap- expenditures. For example, in the United ital investment subsidies, then private sector States, the contribution-financed, universal providers will be unfairly disadvantaged and Medicare program provides limited LTC private sector provisioning of LTC services might be dampened if these subsidies are not 62 See Baeza and Packard (2006) for a discussion on risk- passed on to private providers as well. Often, pooling in general, and Chakraborty, Hofmarcher, and this is not the case, and services procured Koettl (2010) on risk-pooling for LTC specifically. from private providers are reimbursed at 63 See Chakraborty, Hofmarcher, and Koettl (2010) for substantially lower rates. a more detailed discussion. Long-Term Care Expenditures and Policy Options | 53 benefits for the elderly. Any LTC needs ex- services which are excluded from standard ceeding these limited benefits have to be coverage.65 paid by beneficiaries. If beneficiaries cannot Still, nearly all of these instruments afford these services, the tax-financed Med- rely on pay-as-you-go financing mech- icaid program, a social safety net, provides at anisms, which are highly unsustainable least a minimum level of LTC services. Aus- in a rapidly aging society. The urgency tria has a tax-financed universal cash benefit of LTC reform comes from a need to con- that pays benefits based on dependency lev- vince current workers aged 15 to 45 to start els. Yet, these cash benefits are not sufficient saving for their own future LTC needs now, to cover all formal LTC costs, in particular in addition to having to pay for the LTC institutional care, and so the tax-financed needs of their parents’ generation. Ever- social assistance program pays all costs for smaller cohorts of younger people will fol- institutional care that cannot be covered by low these workers. Once today’s young beneficiaries. Germany has a universal, con- workers begin to retire, any system relying tribution-based social LTC insurance, but on inter-generational solidarity will become also uses its social assistance program to pay increasingly unsustainable and will require for LTC costs that cannot be covered from large amounts of debt to be paid off by fu- the by beneficiaries.64 There is evidence that ture generations. tax-financing, in particular via social safety If, on the other hand, private savings nets and infrastructure investments, is being earmarked for the LTC of today’s young increasingly used to finance LTC services. workers were increased, these savings Even in Germany, 20 percent of total expen- could be used to support increased ditures on LTC have been financed via gen- spending on LTC as this large group of eral taxation between 2004 and 2007. cohorts grows older. This implies a buildup Based on successful policies from its of reserves, perhaps through a fully funded European counterparts, Bulgaria could component, and could exist either within so- apply a mixture of instruments to fi- cial insurance or tax-financed funding systems. nance LTC services. It seems that applying If this does not happen, future generations will a mixture of instruments would give the gov- either face much higher contributions and tax ernment additional flexibility to adjust to rates, or LTC benefits will have to be signifi- changing conditions. They range from means- cantly decreased to ensure sustainability. tested to universal entitlements; from privately Yet, given the political and practi- paid or insured to tax and contribution-fi- cal challenges of building up financial nanced benefits and from defined contribu- reserves within a public system, Bul- tion-type to non-defined contribution-type garia should also consider the experi- benefits. Contribution-based benefits have the ence of France. In France, unlike in most advantage of earmarked spending while tax-fi- other European countries, a small private nanced benefits put less of a burden on labor insurance market for LTC provides supple- income. Universal entitlements might have the mentary LTC benefits. The market for pri- benefit of broad political support and may fos- vate supplemental LTC insurance increased ter social cohesion while social safety nets act by 15 percent a year between 2000 and 2007 as benefits of last resort and as an important means of avoiding old-age poverty. Supple- mental and complementary private LTC fi- 64 It should be stressed, though, that safety nets that nancing instruments, in particular LTC support those in need of LTC usually have poverty re- duction as their primary objective, and not providing insurance could serve those who are willing LTC per se. and able to pay more for enhanced LTC ser- 65 For private insurance classification, see Mossialos vices, and for better accommodation and (2002). 54 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria and now covers 4 million people.66 The key ters. These centers could be privately or pub- to the program’s success appears to be its rel- licly owned68 and provide a range of LTC and atively simple design. Benefits are clearly de- rehabilitative services, including home-based fined in cash terms and based on disability and daycare services. levels. This makes the handling of claims Existing community-based LTC much easier than insurance products, which will need to be carefully evaluated be- are based on reimbursements of services (as fore a country-wide rollout. Some mu- in the United Kingdom or the United nicipalities, with the support of NGOs like States). In fact, the French model is based the Red Cross, already seem to have success- more on financial products, like life insur- fully implemented temporary home-based ance, than on health insurance products.67 services or even residential LTC services that Monthly premiums are leveled, but not take care of elderly patients who do not re- guaranteed, and are offered either at an indi- quire hospitalization. vidual or group rate. Eligibility for the ben- With regards to linkages to the health efit is based on total or partial irreversible sector, it is important that the coordina- loss of autonomy, upon which the insurance tion of care services between health and pays a lifetime annuity of €300 to €2,500 the LTC sector focuses on patient needs per month (€600 on average). The uncer- and efficiency gains rather than cost im- tainty about future LTC costs is thus elimi- plications for either the health or the so- nated, and other features that cripple private cial sector. Cost shifting between the sectors LTC insurance in many countries are also should be avoided as much as possible. reduced. In the future, though, as the mar- Because costs for LTC can become ket for these products matures in France, ad- catastrophically high, some form of ditional benefits, including in-kind benefits, risk-pooling is desirable. As pay-as-you- could be added. go financing mechanisms are unsustainable in an aging society, earmarking private sav- ings of young workers for LTC through, for Policy Options example, financial LTC insurance could be an option worth exploring. As demand for LTC services is bound to It is essential that Bulgaria improve increase strongly with population aging, its data collection and recording prac- Bulgaria would need to consider op- tices on LTC services to develop effective tions of scaling up LTC services in a fi- and fiscally responsible policies in this area. nancially sustainable way. Providing better The OECD SHA methodology gives good LTC services would not only protect older guidance on how to record public (and pri- people who become dependent from falling vate) expenditures, and it is important that into poverty, it could also improve the effi- Bulgaria strengthen its efforts along these ciency of Bulgaria’s health sector and reduce lines, also at the local government level. the need for informal care-givers, who are likely to become less available as labor be- comes scarce. Going forward, Bulgaria could thus consider the following policy options: 66 See Le Corre (2008). 67 This product is a typical LTC annuity. For other in- Future investments in LTC and reha- surance products covering LTC risk, see Więckowska bilitative service facilities could convert (2006). municipal hospitals into community cen- 68 Centers could also be run by NGOs. Long-Term Care Expenditures and Policy Options | 55 CHAPTER VI The effect of demographic change on the pension system The future of the Bulgarian pension will impose severe challenges to the pension system largely depends on demograph- system. The population over the age of 65 is ic developments that have already tak- projected to increase by about 60 percent as en place. As Bulgaria’s baby boomers, now a share of the total population between 2011 aged 55–65, start to retire in large numbers, and 2075.70 On top of that, the working-age pension fund revenues will fall while pen- population between the ages of 15 and 64 is sion expenditures rise. At the same time, projected to decline by more than 50 per- Bulgaria’s pension system will experience cent by 2075. The relative size of the elderly the effects of the post-transition demograph- population is also projected to grow faster ic shock that came about as a result of the than the working age population due to low economic turmoil of the early 1990s. In the fertility rates in the past as well as gains in post-transition years, fertility rates in Bul- life expectancy at older ages. Consequently, garia were among the lowest in Europe. Al- the old-age population dependency rate, de- though they have since increased and are fined as the number of elderly per 100 work- currently in line with regional averages, the ing-age persons, is estimated to double from drop led to fewer women of reproductive 25 percent in 2011 to 51 percent in 2075. age today. As a result, even if fertility rates Adding children under the age of 14 would were to increase to replacement levels today, mean that 100 working-age persons sup- the impact would only be felt after the worst ported 77 dependents in 2011. The defini- of the demographic crisis around 2050 oc- tion of working-age population in fact curs. In addition, between 1990 and 2010 understates the severity of the situation as Bulgaria also saw its working-age popula- most people begin employment much later tion contract by about 12 percent. Low than age 15. The dependency rate that best births in the past and high emigration have captures reality is expressed through the sys- led to very few cohorts of 0–20 year olds tem dependency rate, defined as the number joining the pension system today.69 This de- velopment imposes significant challenges to restoring fiscal sustainability. 69 This is also illustrated by the changing shape of the population pyramid for 2011 in Figure VI-1. Dependency rates are expected to 70 The projection period in this chapter is extended to increase markedly over the coming de- 2075 as pension system analysis requires long term cades. Rapid aging of Bulgarian society projections spanning a lifetime of a generation. 57 Figure VI-1: Dependency rates and population pyramid 120 2011 100 91 80 81 71 60 Percent 61 40 51 20 41 31 0 21 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2075 11 Population Dependency Rate System Dependency Rate 1 Old Age Dependency Rate 65 55 45 35 25 15 5 5 15 25 35 45 55 65 Source: United Nations Population projections, NSSI Bulgaria administra- Source: PROST projections based on data from the United Nations tive data. Population Division (2011). of beneficiaries relative to the number of shrinking of the labor force and the established contributors. This statistic stood at 75 elder- difficulty of increasing coverage among the ly or disabled persons per 100 contributors. working-age population. Experience shows It is projected to reach 109 such beneficiaries that over the last decade, the government has per 100 contributors by 2050. The situation on occasion overridden reforms prescribed by is expected to improve after 2050 as the in- law through the implementation of various flow of new retirees decelerates, reflecting discretionary adjustments, including ad-hoc the smaller number of post-transition co- pension increases and Christmas bonuses. For horts retiring (see Figure VI-1). However, at example, despite only partial pension index- that point, 25 percent of the elderly will not ation to wage growth, pension benefits have be eligible for a pension because they will grown in line with wages between 2000 and not have contributed long enough.71 2011 and the ratio of the average old-age pen- The Bulgarian pension system, like sion to the average economy wage has re- many in Europe, will have to adjust to mained relatively constant. these adverse demographic develop- ments. Spending cuts can be achieved either by reducing the number of beneficiaries Key features of Bulgaria’s pension through tighter eligibility criteria, decreasing system the generosity of benefits, or shortening the length of time benefits are given.The menu of Bulgaria’s pension system follows a reform options to increase revenue includes multi-pillar design. Table VI-1 captures increased coverage of the working-age popu- the organization of the Bulgarian pension lation or higher contribution rates. Bulgaria system. It incorporates a mandatory, de- has implemented a mix of these reforms fined-benefit first pillar that’s financed on a through modest increases to the retirement pay-as-you-go basis; a mandatory, fully age, essentially reducing the time spent in re- funded second pillar as well as voluntary ful- tirement, and through changes to the benefit ly funded third pillar. The third pillar has formula and indexation mechanisms to con- low accumulated assets and is therefore of trol pension levels. Despite these reforms, the Bulgarian pension system will likely continue 71They could become eligible for a social pension at to run deficits in view of the projected age 70 if they satisfy a means test. 58 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Table VI-1: Structure of the Bulgarian pension system Zero Pillar Pillar I Pillar II Pillar III Non-contributory Contributory Contributory Contributory Earnings-related (ER) Earnings-related (ER) Earnings-related (ER) Universal social pension Defined Benefit (DB) Defined Contribution (DC) Defined Contribution (DC) State Budget Financed Pay-as-you-go (PAYG) Fully Funded Fully Funded Means tested social pension Mandated (employed, Mandated for individuals born Voluntary self-employed, farmers) after 1959 Reformed in 2000–2003 Reformed in 2000–2003 Introduced in 2002 Introduced in 1999 Source: Authors’ notes. little consequence for the discussion of aver- pension amounts provided from Pillars Zero age-benefit adequacy. The main objective of and I based on years of contributions and in- the second pillar is to enhance benefit ade- come. In 2011 the average old-age pension quacy and, with the exception of the transi- for an individual with a full career, increas- tion costs involved, does not affect public ing to 37 years for women and 40 years for finances. The first pillar matters both in men, amounted to 279 leva or 47 percent of terms of adequacy and public finances and is the average insured wage. Pillar I includes a therefore the focus of this chapter. The sys- minimum pension guarantee that amounted tem also includes a non-contributory zero to roughly 23 percent of the average insured pillar providing a means-tested social pen- wage in 2011. The objective of the mini- sion to individuals above age 70. The struc- mum guarantee is to ensure a pension floor tural reforms that shaped the existing for lower-income earners. pension system took place between 2000 Coverage of the working-age popu- and 2003. Table VI-2 outlines the respective lation is relatively low; currently the Table VI-2: Pension benefit types provided by Pillar 0 and I in 2011 Pillar I Minimum Pillar I Minimum Pension Guarantee Zero Pillar Social Pillar I Pension Guarantee with partial career Pension Age Age: 63m/60f Age: 63m/60f Age: 65 Age: 70 Years of Service 37m/34f 37m/34f Minimum of 15 — Benefit Formula AR*IP*IC*AMII Flat Amount AR*IP*IC*AMII Flat Amount Financing Mechanism Contributory Contributory Contributory Non-Contributory, Means-Tested Pension Amount Average: 279 leva Minimum: 136.06 leva 85 percent of Min. 100.86 leva in 2011 Pension Guarantee, 115.67 leva Percent of average 47 percent 23 percent 20 percent 17 percent insured wage Note: Old age pension benefit formula: Old Age Pension = Accrual Rate * Insurance Period * Individual Coefficient * AMIIa. a Accrual Rate is currently 1.1 percent per insurance year. The Insurance Period consists of contributory and non-contributory periods for which contributions have been paid by the State; the Individual Coefficient is the ratio of an individual’s average insurable income and the national average insurable income, the average is calculated from (i) the best three consecutive years out of the last 15 years of service before 1 January 1997 and (ii) the whole period after 31 December 1996. Currently, the reference period for calculating average insurable income is 18 years, this is gradually extended to the total length of service for persons entering the labor market after 1993. Lastly, the average national monthly insurable income is the income for the previous 12 months before retirement. The effect of demographic change on the pension system | 59 highest potential for expansion is among rent financing gap. The PAYG pension system the youth and persons over the age of promises to support replacement levels that 55. Although coverage rates stood at just cannot be adequately financed by the current above 50 percent in 2011, they are not strik- low contribution rate. Attracting more con- ingly different from other EU10 countries tributors today would certainly result in high- (Figure VI-2a). Coverage rates are indeed er revenues; however in the absence of reform, higher among West European countries, it would also lead to more pension liabilities mostly because of higher statutory retirement when those contributors reach pension age. ages; however none exceeds 80 percent cov- The menu of solutions to this problem entails erage. Coverage is particularly low among the raising contribution rates, raising retirement youth, reflecting high youth unemployment ages or reducing the level of the pension rates (Figure VI-2b). Youth aged of 25–34, benefit. which are most likely to have already com- In an effort to draw in more workers, pleted their education, present the largest op- the government in 2006 reduced contri- portunity for contributor expansion. At the bution rates. The net effect was a sharp de- other end of the age spectrum, there is evi- cline in revenue. While the number of dence that a large number of people still leave contributors did increase during that time, this the labor force before reaching age 65. Post- cannot be solely attributed to the lower con- poning retirement would increase coverage tribution rates, and the higher numbers of among older workers. This is especially true contributors did not compensate for the low- for women whose official retirement age is er contribution rates. Other important factors lower than that of men. included high GDP growth, the echo-boomer The primary objective of pension generation (the children of the baby-boomers) systems is to protect individuals from so- joining the labor market and improved en- cial risks like old age, disability and the forcement efforts. Figure VI-3a compares con- loss of a breadwinning partner. Therefore tribution rates across 28 European countries expanding coverage to include as many people highlighting the particularly low rate in Bul- as possible should be among policymakers’ top garia. This is noteworthy because pension priorities. However, policy makers first need spending levels and replacement rates are com- to evaluate the potential negative impact on parable across these countries and Bulgaria has long-term fiscal sustainability, given the cur- significantly less favorable demographics. The Figure VI-2: Contributors and their age structure in 2011 a. Contributors b. Age Structure of Bulgarian Contributors 90 90 Percent Working Age Population 80 80 70 70 60 60 50 50 Percent 40 40 30 30 20 20 10 10 0 0 Bulgaria Poland Slovakia Latvia Spain Lithuania Slovenia Germany France Switzerland Finland Netherlands Norway Denmark Sweden 18 22 26 30 34 38 42 46 50 54 58 62 Male Female Source: Administrative data and Staff notes. Source: NSSI Bulgaria administrative data. 60 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure VI-3: Pension contribution rates and spending in percent of GDP in 2010 a. Pension Contribution Rates b. Pension Spending in Percent of GDP 40 14 12 30 10 8 Percent Percent 20 6 10 4 2 0 0 Ireland Belgium France Sweden Bulgaria Germany Greece United Kingdom Austria Finland Netherlands Spain EU10 average Italy Portugal Slovakia Lithuania Estonia Bulgaria Czech Republic Romania Latvia Hungary Slovenia Poland EU27 Source: United States Social Security Programs throughout the World, Source: ESSPROS. Europe 2012 total contribution rate, financing both Pillar I Commission 2012a), the PAYG scheme, al- and Pillar II, stood at 17.8 percent in 2011. In- though not fiscally self-sustaining, could dividuals born after 12.31.1959 participate in continue to function as is if the role of the both pillars and contribute 12.8 percent to government in financing pension expendi- Pillar I and 5 percent to Pillar II. Since partic- tures, which corresponds to about 6 percent ipation in the second pillar is only possible for of GDP, remains as significant as it is today. individuals born after 12.31.1959, those born The fact that the system is not self- prior to 1960 contribute the entire 17.8 per- sustaining does not only raise concerns cent to the state pension system.This low pen- about fiscal sustainability, but also equi- sion contribution rate means that the role of ty. The equity of the system of income provi- the government in financing social insurance sion for the elderly will be increasingly has grown to account for almost half of all questioned as today’s near-universal coverage pension spending. Therefore, any further cuts of the elderly will be replaced by a sharply in pension contribution rates should be as- lower coverage rate of the elderly in the fu- sessed very carefully in light of this recent ex- ture. Sizable government subsidies may not be perience and the projected fiscal needs of the of particular concern today, when almost ev- pension system. eryone above the age of 65 gets a pension. The Bulgarian pension system is However, the number of people who have a currently not self-sustaining, relying right to a pension is projected to decline sig- significantly on state subsidies. Employ- nificantly. This will only become apparent as ee and employer contributions financed only the first generation of people who started 53 percent of pension expenditures in 2011 working in the 2000s begins to retire. The (Figure VI-4). The increasing role of the projected decline in elderly coverage is the re- government in financing social insurance sult of high unemployment and sporadic con- through tax revenues raises serious concerns tribution patterns attributed to movement about its financial stability today and in the between formal and informal labor.These cir- future, as our projections confirm. Consis- cumstances create a significant barrier for in- tent with European Commission findings in dividuals to meet old age pension eligibility the 2012 EU Ageing Report (European conditions. Currently, in a given month, the The effect of demographic change on the pension system | 61 Figure VI-4: Financial performance of social insurance 12 100 90 10 80 70 8 60 Percent Percent 6 50 40 4 30 20 2 10 0 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 Pension Exp. %GDP (left axis) Contribution Revenue % GDP (left axis) Contribution Revenue, % Pension Exp. (right axis) Source: NSSI. Bulgarian pension system covers only about takes into account average career earnings. 55 percent of the working-age population. However, the benefit formula also applies an Continuing to subsidize an increasingly small- accrual rate of 1.1 percent per year, scheduled er group of elderly people with state revenue to increase to 1.2 percent in 2017.This accru- finances and excluding a large proportion of al rate translates into replacement rates of 41 the life-long poor raises serious concerns percent for women and 44 percent for men. about the fairness of the system of income With a replacement rate of 44 percent and a provision to the elderly as well as its social sus- system-dependency rate72 of 75 percent, the tainability. In addition, people with pension PAYG pension system would require a pen- rights are likely to be wealthier than those sion contribution rate of 33 percent in order without, as generally people who have worked to be fiscally self-sustainable. Even accounting in the formal sector and have accumulated for the formalized government contribution pension rights tend to have had higher earn- of 12 percent of covered wage bill, the com- ings in their lifetime. bined contribution rate still falls short of clos- Despite a markedly lower pension ing the financing gap. The percentage of contribution rate and worse demo- pension expenditures financed by the state graphics, Bulgaria’s pension system budget grew from about 16 percent in the promises replacement rates of roughly early 2000s to over 50 percent in 2011. As a 47 percent of average wage, in line with result of the mounting financing gap, the Bul- regional averages. Gross replacement rates garian government became a “third insurer” in Bulgaria are similar to those provided by in 2009, contributing 12 percent of the cov- countries experiencing more favorable demo- ered wage bill. Despite the formal introduc- graphics and with significantly higher pension tion of a third insurer, the combined contribution rates (Figure VI-5). Recent re- contribution rate remains insufficient to meet forms to the benefit formula are expected to pension liabilities. The projected replacement control the level of generosity through a clos- level would require an even higher contribu- er link between contributions paid and bene- tion rate in order to attain fiscal balance. fits received. Previously, an individual’s pension was based on earnings averaged over the best three years during the last fifteen prior to re- 72The system dependency rate is defined by the ratio of tirement, while the reformed benefit formula the number of pensioners to the number of contributors 62 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure VI-5: Gross replacement rates 90 80 70 60 50 Percent 40 30 20 10 0 Croatia Sweden Ireland Lithuania Germany Hungary Bulgaria Cyprus Slovakia Estonia Latvia Poland Austria Romania Norway Czech Republic Finland Portugal France Greece Malta Spain Luxembourg Italy Source: Administrative data and staff notes, latest years. Impact of recent reforms pillar included increased eligibility condi- tions for an old age pension, a shift towards Bulgaria’s pension system has under- full-scale inflation indexation of pensions, as gone a number of well-designed sys- well as a comprehensive overhaul of early re- temic and parametric reforms over the tirement programs. The second and third last two decades. The reforms of 2000– pillars were introduced to provide an addi- 2003 transformed the traditional pay-as- tional vehicle for savings and risk-diversifi- you-go system into a multi-pillar structure, cation. While the reforms to the old age including a fully funded mandatory second program were both well designed and effec- pillar as well as a voluntary and fully funded tively implemented, some, such as postpon- third pillar. The motivation behind the ing the retirement age, resulted in increased structural reforms included a weak link be- pressure on the disability program. Disabili- tween contributions and benefits in the pub- ty rolls began to swell as people who wanted lic scheme. The system was providing low to retire but were prevented from doing so and nearly flat-rate pensions; the maximum by the new legislation applied for disability pension benefit amounted to three times the benefits. The flood was stabilized through minimum pension. The average pension stricter eligibility criteria and more effec- benefit was less than 30 percent of the aver- tive enforcement. age wage. The low benefit levels provided Bulgaria recently accelerated re- little incentive for individuals to declare forms to increase the eligibility condi- their earnings and pay into the system. In tions for a pension. Nevertheless, there addition, the emergence of the informal remains a differential between genders economy further enabled individuals to in retirement ages as well as in length of evade pension contributions, rendering the service needed to obtain a pension. The public system largely ineffective. The re- parametric change with the strongest impact forms to the public scheme aimed to allevi- is the increase in the retirement age. By 2017, ate these issues through the introduction of a it will reach 65 for men, up from age 63.8 to- fair and simple benefit formula that provided day and 63 by the year 2020 for women, a a strong link between contributions and slightly larger jump from 60.8 in 2013. Simi- benefits. Additional reforms to the public larly, the length of service required to obtain The effect of demographic change on the pension system | 63 Table VI-3: Summary of main legislated reforms Parameter Pre-2010/2012 reforms Post-2010/2012 reformsa Retirement age 63m/60f 65m/63f Length of service requirement (LOS) 37m/34f 40m/37f Indexation of pensions Swiss Indexation CPI Indexation Contribution rate Pillar I (switchers) 12.8 percent 12.8 percent Contribution rate Pillar II 5 percent 7 percent (as of Jan 2017) Actuarial increase for deferred retirement 3 percent 4 percent Individual pension coefficient calculation period last 16 years Full careerb Accrual rate 1.1 percent 1.2 percent (as of Jan 2017) LOS for military officials and police 25 years 27 years a The December 2010 law would have increased the retirement ages to 65 (men) and 63 (women) by 6 months per year starting in 2021. A new law was passed in January 2012 that accelerates the implementation of the reforms. Under the new law, the retirement age for men and women increased by 4 months - to age 63 and 4 months for men and to age 60 and 4 months for women. From 2013, retirement ages will continue to increase by 4 months each year, until reaching age 65 for men in 2017 and age 63 for women in 2020. b Previously, an individual’s pension was based on earnings averaged over the best three years during the last fifteen years prior to retirement. The reformed benefit formula takes into account the best three years during the last fifteen years until 1996 and average career earnings after this year. a pension will increase to 37 years for wom- informal sector, it is unlikely that Bulgaria en and 40 for men by 2020. A certain per- will be able to finance the same number of centage of the active population will likely years in retirement as other wealthier and fail to accumulate the number of years need- younger European countries. The vast major- ed to qualify for a full old-age pension. Those ity of European countries has already taken individuals would still be able to claim a par- steps towards equalizing retirement ages be- tial contributory pension, contingent upon tween men and women. According to the EU their having contributed to the pension sys- 2012 Aging Report (European Commission tem for a minimum of 15 years. The retire- 2012a), 86 percent of EU27 countries will ment age for a contributory pension with have fully phased in reforms to equalize re- only 15 years of coverage is also being in- tirement ages between genders by the year creased from age 65 to age 67 by the year 2060.The benefits of longer working lives for 2020 for both men and women. women are considerable. Aside from increas- Maintaining a gap between official ing benefit adequacy, more women spending retirement ages for men and women more years in the workforce will lead to more raises both benefit adequacy and fiscal revenue through an expansion of the contrib- sustainability concerns.73 Women tend to utory base as well as lower pension expendi- contribute over a shorter period than men, tures as a result of fewer beneficiaries. partially because they stop working to have Tighter eligibility conditions for children. Enabling women to retire earlier old-age pensions contributed to an in- than men essentially puts them at a higher crease in the disability program. Increas- risk of poverty in old age as fewer years of es to the retirement age and length-of-service contribution translate into lower pension requirements coincided with increased pres- benefits. In addition, women live longer than sure on the disability program. Although it is men. Women spent on average 5 years longer reasonable to observe higher rates of disability in retirement than men in 2010 (Figure VI-6). From a fiscal sustainability standpoint, given the severe projected demographic contrac- 73See Chapter II for a discussion of the labor market tion in Bulgaria as well as the relatively large implications of a gender-differential in retirement age. 64 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure VI-6: Retirement duration in 2010 30 25 20 Years 15 10 5 0 Romania Bulgaria Ireland Hungary Portugal Czech Republic Sweden Slovakia Denmark Netherlands Average Poland United Kingdom Norway Germany Greece Belgium Findland Austria Spain Italy France Male Female Source: European Commission (2012a). among older people, the rapid growth in new and the current difficult labor market condi- pensioners reporting disabilities and claiming tions. International experience shows that benefits in the form of disability pensions is during periods of high unemployment, the likely the effect of individuals seeking alterna- disability program is used as a last-resort safe- tive paths to retirement in view of stricter eli- ty net. When left with no other income pros- gibility conditions in the old age program. pects, many individuals who may have Disability spending as a percentage of GDP previously chosen to work despite a medical more than doubled between 2001 and 2010 issue would now opt for disability benefits. (Figure VI-7) and while disability inflow rates Figure VI-8 illustrates the trend in disability have stabilized since 2006, it is likely that pres- inflow rates and the growth in disability pen- sure on the disability program will intensify, sion spending as percentage of GDP over the given further increases to the retirement age last decade. Figure VI-7: Disability and old age pensioners 2,500 1.8 1.55% 1.6 Spending (in % of GDP) 2,000 1.4 In thousands 1.2 1,500 1.0 0.8 1,000 0.62% 0.6 500 0.4 0.2 0 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Old Age Pensioners (left axis) Disability Pensioners (left axis) Disability Spending, % GDP (right axis) Source: Eurostat. The effect of demographic change on the pension system | 65 Figure VI-8: Projected baseline fiscal PAYG balance (in percent of GDP) 0.0 –1.0 –2.0 –3.0 Percent –4.0 –5.0 –6.0 –7.0 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040 2042 2044 2046 2048 2050 2052 2054 2056 2058 2060 2062 2064 2066 2068 2070 2072 2074 PAYG Balance, %GDP accounting for government contribution (baseline) PAYG Balance, %GDP without government contribution (baseline) Accounting for additional cost for social pensions (without govt. contribution) Accounting for additional cost for social pensions (with govt. contribution) Source: Author’s calculations using PROST. Projections of fiscal balance under projected pension system deficit if a social the baseline pension is provided to individuals who fail to accumulate even the minimum 15 years of Based on PROST projections, Bulgaria’s contributions needed for a partial contributo- pension system is projected to remain in ry pension from Pillar I. As explained above, deficit throughout the projection hori- coverage of the elderly is projected to decline zon. Though current pension spending in in the future.The estimated future coverage of terms of GDP, retirement ages as well as bene- the newly retiring elderly will be about 70 fit generosity are similar to other EU coun- percent, of which an estimated 20 percent will tries, low contribution rates have led to be drawing a partial contributory pension. declining revenues and increasing subsidies The remaining 30 percent will have no right from the state budget. In 2011, employee and to a contributory pension (Table VI-4). Bul- employer contribution financed only 53 per- garia also currently has a Zero Pillar that pro- cent of pension expenditures in 2011. The re- vides a non-contributory, means-tested social maining liabilities were financed from general pension at age 70.The level of the benefit was revenue. Since 2009, the Government has be- set at 100.86 leva in 2011, representing around come a third insurer paying contribution equal 10 percent of GDP per capita (Table VI-4). to 12 percent of the total contributory base. Assuming that all individuals without rights The Government also covers the additional to a pension at age 70 satisfy the income test remaining deficit beyond the 12 percent con- and draw the social pension at a rate equal to tribution. Going forward, the transfers from 13 percent of GDP per capita, the projected the general government budget to the pension cost would gradually begin to grow, reaching fund is projected to decrease from about 6.1 0.5 percent of GDP by 2050 and close to 0.9 percent of GDP in 2012 to about 4.9 percent percent by the end of the projection horizon. of GDP in the medium term before climbing The number of social pension beneficiaries as to about 5.6 percent by 2050. (Figure VI-8). a percentage of the population over the age of Social pension costs are also project- 67 is expected to increase from 0.2 percent to ed to grow. Figure VI-8 also includes the more than 25 percent by 2075. 66 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Table VI-4: Projected pensioners as a percent of population above the retirement age Pension type Age/gender 2013 2020 2030 2040 2050 2075 Full pension from Pillar I Female (age 63) 73 percent 68 percent 62 percent 58 percent 53 percent 51 percent Partial pension (15 yrs Female (age 67) 24 percent 27 percent 25 percent 24 percent 22 percent 19 percent of service) Pillar I No meaningful pensiona Female (age 63) 3 percent 5 percent 13 percent 18 percent 25 percent 30 percent Full pension from Pillar I Male (age 65) 87 percent 68 percent 63 percent 58 percent 53 percent 49 percent Partial pension (15 yrs Male (age 67) 11 percent 25 percent 23 percent 22 percent 20 percent 18 percent of service) Pillar I No meaningful pensionb Male (age 65) 2 percent 8 percent 14 percent 20 percent 27 percent 33 percent a Individuals with less than fourteen years of contributions will receive a negligible pension from Pillar II and no pension from Pillar I. b Individuals with less than fourteen years of contributions will receive a negligible pension from Pillar II and no pension from Pillar I. Pensions benefit levels are projected level from the PAYG scheme alone. to decline relative to the average in- Table VI-5 shows that if an individual was sured wage. A shift towards full-scale in- insured at the minimum wage of 250 leva in flation indexation will gradually result in the 2011 and had worked a full length of service, decline of pension levels relative to average the pension earned would fall below the insured wages. Previously, pensions in Bul- minimum pension guarantee of 136 leva garia were indexed according to the “Swiss within the PAYG pillar. The minimum pen- indexation” rule, i.e. based on 50 percent sion guarantee is only slightly greater than wage growth and 50 percent increase in the partial pension provided with a mini- prices. The projected replacement rate74 for mum of 15 years of contributions. As a re- an existing old-age pensioner, including sult, lower wage earners may find little both the PAYG and funded components is motivation to contribute past the minimum projected to stabilize at around 40 percent 15 years. Additionally, the pension for an av- (Figure VI-9). However, the projected ben- erage wage-earner with a full length of ser- efit from the second pillar assumes a real vice will not be much higher than the growth of 3 percent over the projection ho- minimum guarantee; this induces incentives rizon. Investment risk from the second pillar is fully borne by the individual; therefore it 74The percentage of a worker’s pre-retirement earn- is important to ensure an adequate benefit ings that is paid out by a pension system upon retirement. Table VI-5: Old age pension levels in 2011a When insured When insured at PAYG pension Gender Career Accrual rate Replacement rate at min. wageb ½ insured wagec guarantee Female 34 1.1 percent 44 percent 111 123 136 Male 37 1.1 percent 48 percent 120 152 136 Female 15 Flat 23 percent d N/A N/A 116 Male 15 Flat 23 percent e N/A N/A 116 Source: NSSI Bulgaria a As of 01.01.2017, the second pillar social insurance contribution is expected to be 7% and the accrual rate should be 1.2%, as of 01.01.2017. b Here 250 leva since this was the minimum wage for 2011 the base year for these projections. In 2013, the minimum wage is 310 leva. c Based on current wage income distribution, close to 40 percent of people are insured at income below average insured wage. This trend will lead  to many people receiving pensions below or close to the level of the minimum pension guarantee. d The minimum pension guarantee provided to individuals with full careers amounted to roughly 23 percent of average insurance wage. e The minimum pension guarantee provided to individuals with full careers amounted to roughly 23 percent of average insurance wage. The effect of demographic change on the pension system | 67 Figure VI-9: Projected replacement rate for an average old age pensioner 50.0% 40.0% 30.0% 20.0% 10.0% 2012 2015 2018 2021 2024 2027 2030 2033 2036 2039 2042 2045 2048 2051 2054 2057 2060 2063 2066 2069 2072 2075 Average existing old age replacement rate (Pillar I + Pillar II pension), baseline Average existing old age replacement rate, baseline Source: Author’s calculations using PROST. to under-report wages. By 2060, the average enue added to a smaller earnings-related replacement rate from the PAYG scheme pension financed by private contribution will be only 10 percent higher than the par- would ensure the equity of the pension tial old-age pension provided with 15 years system as coverage among the future el- of service. derly declines. If those individuals who have worked in the formal sector and have made contributions could continue to earn an earn- Alternative financing mechanisms ings-related pension, the consumption-smooth- for a more equitable pension ing element of the system could be preserved. system This arrangement would better promote pov- erty alleviation among the vulnerable elderly The growing role of government in fi- while still providing a mechanism for wealthier nancing social insurance pensions weak- individuals with formal sector employment and ens the link between contribution contribution history to accumulate larger pen- payments and benefits. The state pension sions based on their earnings. Table VI-6 pres- system’s benefits formula was designed to link ents some of the European countries that contributions and benefits. But the presence provide a universal or basic pension. of significant state budget subsidies could re- The starting point of the universal sult in regressive outcomes. One would be the benefit would depend on the subsis- provision of larger benefits to higher-income tence income level of the elderly. The earners who are, to a large extent, financed by process of identifying the level of the benefit general revenue and not solely through pri- would require taking into account all in- vate sector contributions. In addition, such kind benefits and services provided to the el- government transfers could introduce uncer- derly. Such in-kind transfer and subsidies tainty among contributors regarding the sys- could include support with heating pay- tem’s capability to continue providing the ments and free or subsidized healthcare. In promised pension benefits. Lack of confi- addition, many elderly tend to own proper- dence in the system could result in further ty and therefore live rent free. As a result, damage to contribution compliance. careful and detailed assessment of the true A universal benefit for all citizens of needs of the elderly would need to be under- age 65, or higher, financed by general rev- taken in order to arrive at a benefit level 68 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Table VI-6: Countries with universal and basic pensions Universal Pension: all residents or citizens above Basic Pension: Individuals who have fulfilled a certain age a minimum years of service requirement receive a flat benefit, unrelated earnings or the value of their contributions. Benefit may vary based on years of service. Czech Republic, Denmark, Georgia, Ireland, Kazakh- Armenia, Azerbaijan, Croatia, Czech Republic, Ireland, stan, Kosovo, Netherlands, UK Kyrgyz Republic, Lithuania, Luxembourg, Netherlands, Russian Federation, UK high enough to prevent old age poverty but tories will inevitably result in more individu- reasonable enough so as not to create an un- als claiming the minimum pension with manageable fiscal obligation. partial career contributions. A projected 30 The estimated cost of providing a percent of future elderly could claim the so- universal pension to all citizens above the cial pension, a non-contributory means tested age of 67 at a level of 120 leva (equal to benefit provided at age 70, given their income 20 percent of average insured wage) would falls below the minimum income threshold. amount to roughly 2.5 percent of GDP. In 2012, the government contribution of 12 percent of covered wage bill stood at 3 percent Reform scenarios of GDP. Re-allocating the government contri- bution away from social insurance and towards To assess the fiscal impact of different a universal benefit for all elderly would create a reform scenario, this section discusses much more equitable transfer. Separating out three different scenarios that are intended the financing sources would also help create a to address the key challenges of Bulgaria’s more transparent system where earnings-relat- pension systems as in previous sections. The ed pensions are fully financed by private sector scenarios have been modeled using the World contributions. The introduction of a universal Bank’s Pension Reform Options Simulation demogrant pension complemented by a small- Toolkit (PROST). They are as follows: er earnings-related pension would need to be implemented gradually as it would entail a de- • Scenario I: Equalizing retirement ages crease in the accrual rate from Pillar I. for men and women at age 65 The number of people with a re- • Scenario II: Increasing Pillar I contribu- duced pension is projected to increase tion rate by 6 percentage points markedly in the future, primarily as a • Scenario III: Increasing labor force par- result of the currently low contributor ticipation.75 coverage. Challenging labor market condi- tions present a significant hurdle for youth be- tween the ages of 25 and 30 to secure and Scenario I: Impact of equalization of maintain formal employment and to contrib- retirement ages ute to the system on a consistent basis. The presence of a shadow economy also adds to From 2024 to 2050, the equalization of the problem of low contributor coverage. In retirement ages is projected to result in any given month in 2011, roughly 55 percent of the working age population contributed to the pension system. Sporadic contribution 75These projections are based on scenario IV in chap- patterns and shorter than required work his- ter I. The effect of demographic change on the pension system | 69 fiscal savings of about 0.3 percent of ment will result in roughly 1.5 percent higher GDP, eliminating roughly 20 percent replacement rate by the end of the projection of the system’s deficit. Increasing female horizon.This is the result of two extra years of retirement age could mitigate the projected contributions with an annual accrual rate of contraction in the labor force. Labor force 1.2 percent of earnings. Figure VI-10 illus- participation for those over the age of 55 is trates the difference in replacement rates be- particularly low. The projected savings ap- tween the baseline and equal-retirement-age pear modest primarily because the differ- scenario. ence in effective retirement ages between men and women is smaller than the gap be- tween them in official retirement ages. Scenario II: Higher contribution rate Women experience more career breaks than men, for example, to raise children. As a re- A higher contribution rate provides a sult, more women stay longer in the work- clear improvement to the fiscal balance force in order to qualify for pensions. An of the PAYG pension system. Neverthe- even more effective approach to reduce the less, increasing the contribution rate could deficit would be to delay retirement without potentially increase the labor tax wedge as a corresponding adjustment in benefits, es- well as incentives for informal employment. sentially decreasing the accrual rate. As not- Such a reform should, thus, be consid- ed earlier, the effects of an aging population ered carefully. on the pension system can only be addressed through higher contributions, lower benefits or a shorter duration of benefit receipt. Scenario III: Changes in labor force Female replacement rates under the participation equal retirement age scenario are pro- jected to be slightly higher. The average An increase in labor force participation female replacement rate will gradually begin of the overall economy improves the to increase in line with increases to the retire- fiscal balance of the PAYG system. In- ment age. Two additional years in employ- creased labor force participation will result Figure VI-10: Scenario I: PAYG fiscal balance and female replacement rate a. Projected PAYG Fiscal Balance b. Female Replacement Rate 0.0% 40.0% –0.5% 39.0% –1.0% 38.0% –1.5% 37.0% –2.0% 36.0% –2.5% 35.0% –3.0% 34.0% –3.5% 33.0% 2028 2031 2034 2037 2040 2043 2046 2049 2052 2055 2058 2061 2064 2067 2070 2073 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2052 2056 2060 2064 2068 2072 PAYG Balance, %GDP accounting for govt. Female average replacement rate for existing contribution (equal returement age at 65) old age (Pillar I + Pillar II) (equal) PAYG Balance, %GDP accounting for govt. Female average replacement rate for existing contribution (baseline) old age (Pillar I + Pillar II) (baseline) Source: Author’s calculations using PROST. 70 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure VI-11: Scenario II: Projected PAYG fiscal balance 1.0% 0.0% –1.0% –2.0% –3.0% –4.0% –5.0% 2012 2015 2018 2021 2024 2027 2030 2033 2036 2039 2042 2045 2048 2051 2054 2057 2060 2063 2066 2069 2072 2075 PAYG Balance, %GDP accounting for govt. contribution and 6% increase to CR + social pension cost PAYG Balance, %GDP without govt. contribution and 6% increase to CR + social pension cost Source: World Bank Projections based on PROST. Note: Includes the estimated costs associated with the provision of social pensions for individuals without pension rights. Figure VI-12: Scenario III: Projected PAYG balance 0.0% –0.5% –1.0% –1.5% –2.0% –2.5% –3.0% –3.5% 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030 2032 2034 2036 2038 2040 2042 2044 2046 2048 2050 2052 2054 2056 2058 2060 2062 2064 2066 2068 2070 2072 2074 PAYG Balance, % GDP with govt. contribution (baseline) PAYG Balance, % GDP with govt. contribution (higher labor force participation rate in the economy) Source: World Bank Projections based on PROST. Table VI-7: Scenario III: Macroeconomic assumptions High LFP, Baseline TFP 2013 2014 2015 2016 2017 2018 2019 2020 2025 2030 2040 2050 2055 2065 2075 Real GDP 3.0 3.0 3.5 3.2 2.7 2.8 2.8 3.3 3.2 2.5 1.5 0.9 0.9 0.9 0.9 Growth ratea Real Wage 3.2 3.0 2.9 3.2 3.5 3.3 3.3 3.0 2.8 2.2 2.2 1.8 1.8 1.8 1.8 Growth ratea a In percent. in slightly higher GDP. This scenario does individuals entering the labor force would not assume an increase in the number of engage in formal labor and begin to contrib- people contributing to the pension system, ute to the pension system. Also as explained because it is difficult to predict how many earlier, an increase in the number of The effect of demographic change on the pension system | 71 contributors would result in fiscal savings at equality. The fact that the system is not first but would ultimately result in more self-sustaining does not only put pressure on pension expenditures in the absence of re- government debt but raises significant con- forms. The PAYG system is projected to re- cerns about equity as the government will main in deficit; bringing more people into it continue to subsidize an increasingly small will likely further exacerbate the fiscal prob- group of relatively well-off elderly while ex- lem. Moreover, the government has official- cluding a large portion of the life-long poor. ly assumed the role of a third insurer paying In view of changing demographics and diffi- at the rate of 12 percent of the covered wage cult labor market conditions, the financing bill. If the latter were to increase the govern- mechanism for Bulgaria’s pension system ment’s liabilities to the pension system would may need to be reviewed so that the contrib- also raise. utory system can become self-sustaining, that it is financed solely by private sector contri- butions. Policy options Equalizing retirement ages for men and women at age 65, while preparing In the absence of reform, Bulgaria’s pen- further increases to the retirement age sion system is projected to remain in beyond age 65 would be important to deficit. Low contribution rates have led to consider 76 . Women tend to contribute for a declining revenues and increasing subsidies shorter periods than men, which is to some from the state budget in recent year. In 2011, extend the result of career interruptions ex- 47 percent of pension expenditure was fi- perienced during childbearing years. En- nanced from general revenues, corresponding abling women to retire earlier than men to about 6 percent of GDP. Going forward the raises concerns about the adequacy of their transfers from the general government budget pension benefits. In addition, an earlier re- to the pension fund is projected to decline to tirement age for women, who have a longer about 4.9 percent by 2030—due to declining life expectancy than men, further increases coverage—before reaching 5.6 percent in the period of benefit receipt for women rel- 2050. Increasing the contribution rate levied ative to men. on individual wages by 6 percentage points Strengthening the disability certifi- clearly improves the fiscal balance. However, cation processes could help mitigate an such a decision would require careful consid- increase in disability claims stemming eration for the potential negative effects on the from high unemployment and stricter labor market and overall economy. In the con- eligibility rules in the old age program. text of a rapid demographic contraction pur- Disability spending as a percentage of GDP suing both objective of poverty alleviation and more than doubled between 2001 and 2010. income replacement inevitably leads to the Although the inflow rates were stabilized in compression of the income distribution of the 2006, it is likely that the pressure on the elderly and reduced incentives to contribute. An increase in the retirement age is the main 76 One way to address the risk of longevity is to link the tool to mitigate this trend. Nevertheless, Bul- statutory pension age to life expectancy. Fourteen EU garia could consider a smaller increase in the Member States have already introduced a variety of PAYG contribution of around 2 percent that sustainability factors and/or other reduction coeffi- could still strengthen the fiscal state of the first cients that ultimately determines the amount of the pillar without introducing major labor mar- pension entitlement. For example, Italy and the Czech Republic have linked the retirement age to increases in ket distortions. life expectancy, essentially removing the need for fur- The prevailing issue with the Bul- ther reforms to the pension age (European Commis- garian pension system is its inherent in- sion, 2012) 72 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria disability program will intensify if the re- financing needs and ultimately jeopardize tirement age increases amid difficult labor the long-term fiscal sustainability of pension market conditions. the system. Pension reform is inherently a Avoiding further ad-hoc pension difficult undertaking from a political stand- increases would facilitate long-term point, especially when pension reforms be- fiscal planning. Evidence shows that in- gin to exhibit an expenditure cutting impact. dexation reforms did not fully hold and re- Nevertheless, policymakers should refrain current increases in pensions above those from resorting to ad-hoc measures in order prescribed by the law occurred. Such prac- to ensure the fiscal sustainability of the pen- tices make it difficult to accurately project sion system in the long-term. The effect of demographic change on the pension system | 73 CHAPTER VII Macro-Fiscal Implications of Bulgaria’s Demographic Change Bringing together the analyses of the transfers to the pension fund and student previous chapters, this chapter quantifies numbers, all reflecting Bulgaria’s specific cir- the macro-fiscal impact of Bulgaria’s cumstances. The model does not attempt to demographic change over time in a co- provide a novel contribution to the generic herent setting. A simple modeling platform debate of the impact of demographic change brings together independent projections of by solving individual optimization problem the population, the labor force by age and ed- on the basis of a labor-leisure choice, and ab- ucation level, government transfers to the stracts from characterizing the impact of gov- pension fund and student numbers as well as ernment expenditures on household welfare. estimates of public expenditure-age profiles Simulations suggest a significant for health and long-term care.77 Since the im- slowdown in Bulgaria’s growth and a pact of demographic change on key expendi- steep increase in its debt-to-GDP ratio. ture items, such as health-care, education and Under the baseline projections, GDP growth pension can be compounding and offsetting, reaches 2.6 percent in 2016. It then decreases it is important to evaluate the fiscal impact of gradually to about 1 percent in the early demographic change in a coherent frame- 2040s, and continues to decline to 0.7 per- work. This model extends existing analyses of cent by the end of 2050, the projection hori- the long-term impact of Bulgaria’s demo- zon. Public expenditures for health-care are graphic change (see, for example, European expected to increase from 4.1 percent of the Commission 2012b) by combining detailed GDP in 2012 to 5.1 percent by 2050. Long- sector analyses in a consistent framework, term care expenditures are expected to rise which feeds into a fiscal sustainability analysis. from 0.45 percent of GDP to 0.76 percent by The model also assesses the impact mid-century. Government transfers to the of policy changes through alternative pension fund decrease from 6.1 percent of scenarios. In particular, it explores how GDP to 4.9 percent in the medium term and changes in net migration and life expectancy, increase to about 5.6 percent by 2050. Public LFP rates, the retirement age, productivity growth and public investment in education 77See Chapter I for a discussion of population projec- would alter the baseline findings. These alter- tions and public expenditure-age profiles, and Chapter native scenarios are based on consistent pro- VI for projections of government transfers to the pen- jections of the labor force, government sion fund. 75 Table VII-1: Baseline Assumptions 2012 2015 2020 2025 2030 2035 2040 2045 2050 (Percent) Baseline TFP Growth Rate 0 2 1.3 1.2 1.2 1.2 1.2 1.2 1.2 Global Interest Rate (Before Tax) 4.5 4.5 4.5 4.5 4.5 4.5 4.5 4.5 4.5 Interest Rate Differential 2.5 2.5 2.3 2.2 1.7 1.2 0.6 0.3 0.1 Unemployment Rate 12.3 11.1 9.1 7.1 5.8 5.3 5 5 5 (Percent of GDP) Direct Tax Revenues 5.4 5.5 5.9 6.1 6.3 6.4 6.4 6.5 6.5 Indirect Tax Revenues 17.2 16.4 16.2 16.2 16.2 16.2 16.2 16.2 16.2 Non-Tax Revenues 4.8 3.8 3.6 3.6 3.6 3.6 3.6 3.6 3.6 Grants 3.1 3.8 3.4 3.3 3.2 3.1 3.1 3 3 Contribution to EU Budget 1 1.1 1.1 1.1 1.1 1.1 1.1 1.1 1.1 Other Expenditures 18.6 19.5 19.4 19.2 19.1 19.1 19 19 19 expenditures on education are projected to Estimating the capital stock in decline from 4.6 percent of GDP in 2012 to Bulgaria is a challenge. In many transition 3.2 percent by 2050 as the number of stu- economies, a large unknown part of the cap- dents is projected to shrink by nearly 50 per- ital stock has been destroyed as factories were cent. The assumptions underlying these abandoned and equipment was scrapped in projections are quite conservative. Still, Bul- the transition period. Moreover, the intro- garia’s ratio of public debt to GDP is expect- duction of new technologies by the devel- ed to increase from about 18 percent to 51 oped economies accelerated the depreciation percent by the end of projection period. of existing capital (Campos and Coricelli 2002). Estimates suggest that these losses are likely to be large.79 Relying on a standard Key baseline assumptions perpetual inventory approach is thus likely to lead to an overestimation of the capital stock The baseline is designed to reflect the in Bulgaria. For example, based on this ap- macro-fiscal outcomes in the absence proach, Ganev (2005) estimates a capital out- of policy adjustments. The demographics put ratio of 3.18 for Bulgaria in 2007. This under the baseline scenario are based on model uses a more conservative estimate of United Nations population projections. The 2.55 in the base year, which is in line with the labor force projections are based on ILO capital-output ratio for 2010 discussed in projections till 2020 and assume constant la- IMF (2010) and the EU estimate for Slovenia bor force participation rates by gender and (European Commission 2011). five-year age groups thereafter.78 Total factor The model assumes that Bulgaria is productivity (TFP) is assumed to grow at 1.2 a small and open economy. This assump- percent in the long term, which is slightly above the average TFP growth projected for the EU (European Commission 2012a), in 78 For a detailed discussion of different labor force sce- narios, see Chapter I. line with a gradual convergence in the ab- 79 Deliktas and Balcilar (2005), for example, estimate sence of productivity-enhancing policy re- that up to 50 percent of the communist-era capital forms. stock has been destroyed in the early transition period. 76 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure VII-1: Bulgaria’s capital flows in percent of GDP 30% 30% 30% 20% 2001–04 20% 2005–08 20% 2009–11 10% 10% 10% 0% 0% 0% –10% –10% –10% –20% –20% –20% Bulgaria EU10 EU15 Bulgaria EU10 EU15 Bulgaria EU10 EU15 FDI Portfolio Other CA Balance Total Capital Flows Source: Authors’ calculations based on Eurostat. tion is motivated by the fact that Bulgaria has Projecting the public health-care attracted exceptionally large capital flows of expenditures is a daunting task. Public up to 30 percent of GDP in net terms prior health-care expenditures are projected on the to the 2008 financial crisis. Even after the cri- basis of age-specific health-care expenditures sis, Bulgaria’s capital flows remain high com- that are indexed to per-capita income through pared to other EU countries (Figure VII-1). income elasticity.81 This elasticity is assumed In an open economy environment, capital to be slightly larger than one (1.15), reflecting flows respond to the differences between the a catch-up effect of public health expendi- domestic economy and the rest of the world tures given the fact they are currently low in terms of risk-adjusted returns on invest- compared to the rest of the EU. In order to ment. Between 2012 and 2050, the interest calculate the total public health-care expen- rate differential between the Bulgarian rate diture share in GDP, these individual expen- and the rest of the world’s is assumed to de- ditures are multiplied by the average cline from 2.5 percentage points in 2012 to per-capita income and the number of indi- 0.1 percent in 2050, which encourages fur- viduals in each five-year group for a given ther capital inflows. This path is motivated by year. The projection methodology relies on several factors including a declining risk aver- an implicit assumption: health-care consump- sion of global investors as well as economic tion remains proportional across age groups and institutional convergence of the Bulgari- as life expectancy increases. This is a very an economy with the rest of the EU.80 On the conservative assumption. In Bulgaria, for ex- other hand, the decline in Bulgaria’s labor ample, public health-care expenditures of force depresses the return on capital, discour- 70–75-year-olds are about 3 times higher aging additional inflows. than those of 50–55-year-olds. In many EU15 The unemployment rate is assumed countries, this ratio tends to be above 5 (Eu- to reach 5 percent just before 2040. Since ropean Commission 2009). the beginning of the new millennium, Bul- garia’s unemployment rate has steadily de- clined from around 20 percent to 5.6 percent 80 The model does not enable to quantify the impact of by 2008. In the wake of the 2008 financial aging on domestic savings. However, as discussed in crisis, unemployment has increased steeply, Chapter I, this impact depends on a variety of factors, including the demographic drivers of aging and char- reaching 12.3 percent in 2012. As the econo- acteristic of the social security system. my converges to its long-term potential, so 81 See Chapter I for a discussion of the public expendi- does the unemployment rate. ture-age profiles for health and long-term care. Macro-Fiscal Implications of Bulgaria’s Demographic Change | 77 Long-term projections of revenues medium term, employment is projected to are based on the following assump- contribute positively to growth. This is driv- tions: Non-tax revenues and indirect tax en by the fact that the assumed reduction in revenues are assumed to have unitary in- the unemployment rate in the medium term come elasticities; therefore, their share re- has a positive effect on employment, which mains constant in GDP in the long term. initially outweighs the labor-force decline. The only adjustment comes from small re- Once the unemployment rate reaches its ductions, between 2012 and 2015, of indi- long-term equilibrium of 5 percent, only rect revenues from 17.2 percent to 16.4 the negative effect of a decreasing labor force percent and from non-tax revenues from 4.8 remains (Figure VII.1, Panel (b)). However, percent to 3.8 percent, which are in line TFP growth and increasing capital intensity with the Medium Term Fiscal Framework raise the productivity of labor throughout of the Government of Bulgaria (2012). Di- the projection horizon. On average, the rect tax revenues are, on the other hand, GDP per worker grows by about 2.4 percent projected to exhibit modest improvement per year. over time, increasing from 5.4 percent of Public health-care expenditures as GDP in 2012 to 6.5 percent in 2050, repre- a share of GDP rise throughout the senting better tax collection. Box VII-1 pro- projection period. They climb from 4.1 vides more information about the underlying percent in 2012 to more than 5.1 percent by structure of the model. 2050 (Figure VII-2, Panel d). This increase arises from a combination of two factors: first, public health-care expenditures per Key Findings person are higher for the elderly. An in- crease in the average age of society, thus, in- Baseline Results creases the share of aggregate income spent on health-care. Second, assuming that Bul- In the long run, GDP growth is pro- garia’s public health-care expenditures as a jected to decline. Under the baseline pro- share of GDP will eventually catch up with jections, annual GDP growth reaches 2.6 other European countries83, health-care percent in 2016. It then decreases gradually consumption is assigned an income elastici- to about 1 percent in the early 2040’s, and ty slightly greater than one (1.15). As a re- continues to decline to 0.7 percent by the sult, even after controlling for the change in end of the projection horizon. As a result, the age composition, public health-care ex- real GDP is projected to grow from about penditures still increase as a share of GDP. US$77 billion in 2012 to US$126 billion in Long-term care expenditures, which are 2050. Monte Carlo simulations show that, in classified under the other expenditures cate- the long term, the growth rates are expected gory, are projected to increase from 0.45 to remain within a band of 2.2 percent and percent in 2012, to 0.76 in 2050. These pro- 0.2 percent with a 90 percent degree of con- jections are in line with European Commis- fidence.82 The growth rate in terms of GDP per capita, on the other hand, reaches a peak 82 These simulations characterize shocks to TFP growth of 3.3 percent in 2016, and then gradually that are identically and independently distributed with slows down to 1.5 percent by the end of the zero mean and 0.75 percent standard deviations. projection horizon. 83 Bulgaria’s public health expenditures as a share of This decline is largely driven by the GDP are very low compared to other European coun- tries even when taking into account income levels. As shrinking work force. From 2021 on- discussed in Chapter IV, it is reasonable to assume the wards, employment is projected to depress public health expenditure per capita as a share of GDP growth as the labor force declines. In the will rise. 78 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Box VII-1: Underlying models The analysis in this chapter draws on simulations from a projection model by Onder, Pestieau, Ley (2013), and analytical findings from an Overlapping Generations (OLG) model developed by Dedry, Ley, Onder, and Pestieau (2013). The first model characterizes Bulgaria as a small and open economy, where capital is the mobile factor. Output is produced by a stan- dard Cobb-Douglas production function with constant returns to scale and two factors of production: capital and efficiency unit of labor. Physical labor is converted to efficiency units by using a Mincerian transformation. Labor supply is exogenous to the model, and the equilibrium level of capital is given by equalization of risk adjusted returns between domestic markets and the rest of the world. In the model, government collects taxes, receives non-tax revenues and grants, and provides public services. Tax revenues comprise indirect taxes, capital income taxes, and labor income taxes, whereas public expenditures include education, health, pension fund transfers, and other expenditures. The OLG model uses a two period environment with endogenous age of retirement. An interesting finding is that the impact of aging on domestic savings and, hence, on investment and capital accumulation, depends on the drivers of aging and the prevailing social security system. Aging, which is driven by a decline in fertility unambiguously fosters savings, whereas an increase in longevity (or decrease in mortality) in a society with flexible retirement age and pay-as-you go social security has ambiguous effects on savings. Even though an increase in savings would be expected under reasonable assumptions, if the pensions system is rather generous, then a decrease in savings is also possible. Moreover, the precise type of financing also affects the impact of aging on capital accumulation. In the case of a defined contribution system, where the current beneficiaries bear the burden, a reduction in fertility rates increases the savings and the capital accumula- tion. If the benefits are defined, then the net effect is ambiguous. Inputs Model Mechanics Results Pension Fund (PRQST) Population Debt Main Inputs Real Sector Fiscal Sector Dynamics Dynamics Indicators Scenarios Scenario Analysis sion (2012a) projections, but ignore the fact increasing steeply again in the longer that public LTC expenditures are likely to run. The transfers from the general govern- be underestimated and that the public pro- ment budget to the pension fund take two vision of LTC services in Bulgaria is cur- forms in Bulgaria. First, an amount that is rently likely to be sub-optimal (see Chapter proportional to the total current pension V for a detailed discussion). contributions from employee payrolls is trans- Fiscal pressures arising from the ferred from the general budget to the pension deficits in the social security system de- fund, as enacted by law. Second, if the pension cline in the medium-term before fund runs a deficit within a year, the gap is Macro-Fiscal Implications of Bulgaria’s Demographic Change | 79 closed by a second-tier transfer from the gen- Figure VII-2, the sum of these two types of eral government budget within the same year. transfers exhibits a small decrease from about As shown in Panel (d) of Figure VII-2, the 6.1 percent of GDP in 2012 to about 4.9 per- sum of these two types of transfers exhibits a cent in the medium term, before climbing to small decrease from about 6.1 percent of about 5.6 percent by the end of the projec- GDP in 2012 to about 4.9 percent in the me- tion horizon, mainly driven by the number dium term, before climbing to about 5.6 per- of contributors and beneficiaries.85 cent by the end of the projection horizon, A decline in the number of students mainly driven by the number of contributors is expected to drive public education and beneficiaries.84 expenditures down. The total number of A decline in the number of students students was 1.03 million in 2012; this num- is expected to drive public education ber is projected to decline to 527,000 in expenditures down. The total number of 2050. Although education expenditures per students was 1.03 million in 2012; this num- student are assumed to grow at the same rate ber is projected to decline to 527,000 in as GDP per capita, the sheer size of the re- 2050. Although education expenditures per duction in enrollment reduces public educa- student are assumed to grow at the same rate tion expenditures. As a result, public as GDP per capita, the sheer size of the re- education expenditures will decline from 4.6 duction in enrollment reduces public educa- percent of GDP in 2012 to 3.2 percent by tion expenditures. As a result, public 2050. These projections implicitly assume education expenditures will decline from 4.6 that there is no need for “protected” educa- percent of GDP in 2012 to 3.2 percent by tion institutions and additional investments 2050. These projections implicitly assume in life-long learning. Moreover, increasing that there is no need for “protected” educa- the quality of education that is crucial for tion institutions and additional investments sustaining productivity growth may require a in life-long learning. Moreover, increasing more-than-proportionate increase in the the quality of education that is crucial for per-student expenditures, which would tilt sustaining productivity growth may require a up the downturn trajectory of education ex- more-than-proportionate increase in the penditures (see Chapter IV for a detailed dis- per-student expenditures, which would tilt cussion). up the downturn trajectory of education ex- Bulgaria’s demographic transition penditures (see Chapter IV for a detailed dis- is projected to lead to a steady increase cussion). in public debt. The fiscal budget is pro- Fiscal pressures arising from the jected to incur primary deficits throughout deficits in the social security system the projection horizon, which are expected decline in the medium-term before in- to be close to 1 percent in the longer term creasing steeply again in the longer (Panel (e) in Figure VII-2)86. As a result, the run. The transfers from the general govern- ment budget to the pension fund take two forms in Bulgaria. First, an amount that is 84 For a detailed discussion of the evolution of Bulgar- ia’s PAYG balance and assumptions underlying these proportional to the total current pension simulations, see Chapter VI. contributions from employee payrolls is 85 For a detailed discussion of the evolution of Bulgar- transferred from the general budget to the ia’s PAYG balance and assumptions underlying these pension fund, as enacted by law. Second, if simulations, see Chapter VI. 86 The most recent budget projections from the Minis- the pension fund runs a deficit within a year, try of Finance of Bulgaria report a 0.2 percent prima- the gap is closed by a second-tier transfer ry surplus in 2012, 0.3 percent deficit in 2013, 0.2 from the general government budget within percent deficit in 2014, and a balanced primary ac- the same year. As shown in Panel (d) of count in 2015. 80 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure VII-2: Fiscal Projections under the Baseline Assumptions a. GDP Growth b. Decomposition of GDP Per Capita Growth 0.06 4% 0.05 3% 0.04 Contribution to GDP GDP Growth Rate per Capita Growth 0.03 2% 0.02 0.01 1% 0 0% –0.01 –0.02 –1% 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2050 66.7 percent 75 percent 90 percent Time Mean 99 percent 95 percent Growth Accounting Total Factor Productivity Population Human Capital Capital Stock Employment Ratio c. Revenue Decomposition d. Age Related Public Expenditures 35 6.5 30 Age Related Expenditures 6.0 Revenue Composition (Percent of GDP) 25 (% of GDP) 5.5 20 5.0 15 4.5 10 4.0 5 3.5 0 3.0 2010 2015 2020 2025 2030 2035 2040 2045 2050 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2050 Time Time Revenues Age Related Expenditures Direct Tax Revenues Non-Tax Revenues Education Expenditures Health Care Expenditures Health Care Contributions Indirect Tax Revenues Human Transfers to the Pension Fund Non-Pension Social Contributions Grants e. Primary Balance f. Debt to GDP Ratio 0.005 0.60 Primary Balance to GDP Ratio 0.55 0 0.50 Debt to GDP Ratio 0.45 0.40 –0.005 0.35 0.30 –0.01 0.25 0.20 –0.015 0.15 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 66.7 percent 75 percent 90 percent 66.7 percent 75 percent 90 percent Mean 99 percent 95 percent Mean 99 percent 95 percent Source: Authors’ simulations. Macro-Fiscal Implications of Bulgaria’s Demographic Change | 81 Table VII-2: Decomposition of Debt Dynamics under the Baseline Scenarioa 2015 2020 2025 2030 2035 2040 2045 2050 Change in Debt to GDP Ratio 1.2% 0.8% 0.7% 0.6% 0.6% 0.7% 1.0% 1.1% Growth Effect –0.5% –0.5% –0.4% –0.4% –0.4% –0.4% –0.4% –0.3% Interest Payments 0.8% 0.4% 0.3% 0.4% 0.4% 0.4% 0.5% 0.5% Primary Balance 0.8% 0.9% 0.8% 0.7% 0.6% 0.7% 0.9% 0.9% a Annual contributions in percentage points, 2012–2050. debt-to-GDP ratio increases from its current higher investment.88 Two alternative TFP level of about 18 percent to about 51 percent scenarios have been developed to explore by 2050. Even though this shows a substan- the impact of different TFP assumptions on tial deterioration, this level is still within the the projection results. The low TFP growth limits defined by Maastricht criteria. How- series in Panel (a) in Figure VII-3 is 0.15 per- ever, the protracted upward trend is cause cent smaller than in the baseline case in the for concern. long term. In the medium term, differences Protracted primary deficits are the range from 0.8 percent in 2016 to 0.18 per- most prominent factor leading to an ac- cent in 2021. The low TFP growth series cumulation of public debt over the pro- approaches the average TPF growth of EU jection period (Table VII-2). Primary countries projected by the European Com- deficits increase the ratio of debt to GDP by mission (2012a).89 The high TFP growth se- about 0.9 percentage points per year. Interest ries is about 0.3 percentage points higher payments raise the ratio by about 0.4 percent- than the baseline in the long term and is age points per year, roughly offsetting the similar to the Bulgaria-specific projections GDP growth effect.87 in the European Commission (2012a) re- port. In this case, GDP per worker grows by an average of 2.9 percent per year. Alternative Scenarios Public health-care expenditures are higher in a richer society. An interesting In order to explore which policies can result of higher productivity is that the share mitigate the negative consequences of of GDP that is spent on public health-care Bulgaria’s demographic change, three increases slightly compared to the baseline. alternative classes of scenarios are presented Public health-care expenditures were assumed here affecting: i) total factor productivity, to grow by a rate proportional to GDP per ii) labor force, and iii) education of the pop- capita. In an economy with a shrinking ulation. 87 Since the model does not capture inflation or ex- change rate movements there are no other valua- Total Factor Productivity Scenarios tion effects. 88 Labor supply is assumed to be inelastic in the model and thus does not change under the different TFP The assumption of high productivity growth scenarios. growth directly raises GDP growth. 89 Differences between the two series mainly lie in medi- But there is also an additional indirect um term dynamics. The EC study employs a relatively effect. When TFP increases, marginal pro- stable TFP growth trajectory, whereas this model incor- porates the current economic downturn and projects a ductivities of production factors increase as recovery in the medium term. This leads to first lower well. In response, there will be further capi- and then higher TFP growth rates relative to the EU av- tal inflows into an open economy and thus erage in the short and medium term, respectively. 82 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure VII-3: Projections under TFP scenarios a. TFP Growth Rate b. GDP Growth 0.015 0.035 0.030 0.010 GDP Growth Rate (%) TFP Growth Rate 0.025 0.005 0.020 0.000 0.015 0.010 –0.005 0.005 –0.010 0.000 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 Baseline TFP Growth Low TFP Growth Baseline TFP Growth Low TFP Growth High TFP Growth High TFP Growth c. Public Expenditures for Healthcare d. Transfers to the Pension Fund 5.4 6.2 Public Healthcare Expenditures Transfers to the Pension Fund 5.2 5.0 5.8 (% of GDP) 4.8 (% of GDP) 5.4 4.6 4.4 5.0 4.2 4.0 4.6 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 Baseline TFP Growth Low TFP Growth Baseline TFP Growth Low TFP Growth High TFP Growth High TFP Growth e. Primary Balance f. Debt to GDP Ratio 0.2 0.65 0.0 0.55 Debt to GDP Ratio Primary Balance 0.2 (% of GDP) 0.4 0.45 0.6 0.35 0.8 0.25 1.0 1.2 0.15 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 Baseline TFP Growth Low TFP Growth Baseline TFP Growth Low TFP Growth High TFP Growth High TFP Growth Source: Authors’ simulations. Macro-Fiscal Implications of Bulgaria’s Demographic Change | 83 Figure VII-4: Projections under LFP scenarios a. Total Labor Force b. GDP Growth 3.6 3.5 3.4 3.0 Labor Force (Millions) GDP Growth Rate (%) 3.2 2.5 3.0 2.0 2.8 1.5 2.6 2.4 1.0 2.2 0.5 2.0 0.0 2024 2032 2036 2020 2028 2012 2016 2040 2044 2048 2024 2032 2036 2020 2028 2012 2016 2040 2044 2048 Baseline LF High LFP Scenario Baseline LF High LFP Scenario Retirement Age Scenario EuroPop Projections Retirement Age Scenario EuroPop Projections c. Public Expenditures for Healthcare d. Transfers to the Pension Fund 5.40 6.1 Public Healthcare Expenditures Transfers to the Pension Fund 5.9 5.20 5.7 5.00 (% of GDP) 5.5 (% of GDP) 4.80 5.3 4.60 5.1 4.40 4.9 4.20 4.7 4.00 4.5 2024 2032 2036 2020 2028 2012 2016 2040 2044 2048 2024 2032 2036 2020 2028 2012 2016 2040 2044 2048 Baseline LF High LFP Scenario Baseline LF High LFP Scenario Retirement Age Scenario EuroPop Projections Retirement Age Scenario EuroPop Projections e. Primary Balance f. Debt to GDP Ratio 0.4% 0.55 0.50 Primary Balance (% of GDP) 0.2% 0.45 Debt to GDP Ratio 0.0% 0.40 –0.2% 0.35 –0.4% 0.30 –0.6% 0.25 –0.8% 0.20 –1.0% 0.15 2024 2032 2036 2020 2028 2012 2016 2040 2044 2048 2024 2032 2036 2020 2028 2012 2016 2040 2044 2048 Baseline LF High LFP Scenario Baseline LF High LFP Scenario Retirement Age Scenario EuroPop Projections Retirement Age Scenario EuroPop Projections Source: Authors’ simulations. 84 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria population, GDP per capita grows faster than ing of the labor market, education and train- GDP itself. In addition, when the income ing and the health services which have been elasticity of health expenditures is greater discussed in detail in previous chapters, these than one, the differential between GDP per policies could also take the form of establish- capita growth and GDP growth has an even ing a more predictable business environment, more pronounced impact on public expendi- improving innovation and policies and pro- tures. Higher TFP growth magnifies this viding better basic road infrastructure and effect, boosting public expenditures even as a improved performance of the energy sector.90 share of GDP. Higher productivity growth helps reduce the deficit of the social security Labor Force Participation and system. As Panel (d) shows, high TFP Population Scenarios growth leads to smaller transfers. Higher TFP growth generates a nominally greater Since the evolution of the labor force pension fund deficit that needs to be fi- will have a significant impact on Bul- nanced by transfers from tax revenues. This garia’s growth outlook, the model sim- is because benefits are indexed to wage ulates the impact of different labor force growth, and higher productivity growth in- and population projections. Alternative creases real wages faster. However, higher labor supply scenarios presented here are the productivity also increases GDP, and in this retirement-age scenario, which shows the case its impact on the GDP is comparatively impact of current legislated reform of an in- larger than its impact on the pension fund crease in retirement ages of men and wom- deficit. Therefore, the amount of transfers as en91 and the high LFP scenario. The latter a share of GDP is smaller in the case of a assumes a convergence in LFP rates to high high TFP growth scenario, compared to the LFP countries (Scenario II in Chapter I).The baseline scenario. A similar reasoning ex- scenario implies LFP beyond 80 percent even plains why a lower TFP growth implies a among those aged 70 to 74. The scenario can greater burden for the fiscal budget levied by be interpreted as an upper bound.92 the social security system. In order to capture the impact of a In the long run, maintaining a high reduction in net emigration, this chapter level of productivity growth is crucial also presents results based on Eurostat for fiscal sustainability. A high TFP growth population projections EUROPOP2010. reduces the primary deficit as displayed in While the UN extrapolates current immigra- Panel (e) and thereby the debt-to-GDP ratio from 51 percent under the baseline to 45 per- 90 As mentioned in the introduction the discussion of cent in 2050. Under the low TFP scenario, productivity-enhancing reforms through channels the same ratio increases to about 64 percent other than improving education and health will be rel- over the projection horizon. In the absence of egated to a second study. a policy change, this result suggests that if the 91 A law passed in January 2012 raised the retirement Bulgarian economy does not outperform the age for men and women by 4 months to age 63 and 4 months for men and to age 60 and 4 months for wom- EU average in terms of TFP growth, there en. From 2013 retirement ages will continue to in- will a substantial risk in the long run of crease by 4 months each year, until reaching age 65 for breaching the Maastricht criterion, which men in 2017 and age 63 for women in 2020. limits the public debt to 60 percent of GDP. 92 As discussed in Chapter I, this scenario is likely to re- An essential point here is that maintaining flect combination of ambitious reforms, including changes in tax treatment of second wage earners, child high productivity growth is not an automatic care subsidies, increases in the statutory retirement age process. It entails active policies and planning. for men and women and health sector reforms which In addition to improvements in the function- support the extension working lives. Macro-Fiscal Implications of Bulgaria’s Demographic Change | 85 tion trends and projects 10,000 net emigrants as a share of GDP decrease at the same time per year throughout the projection period, due to a higher number of contributors; low- Eurostat assumes an increase in net emigra- er growth of real wages in relative terms, tion from 10,000 persons per year to 14,000 which is used as a base for pension benefits; in 2020, which then declines steadily at the and a higher denominator. The high labor- beginning of the 2030s for the EURO- force scenario leads to a strong improvement POP2010 projections. Between 2030 and in primary balances (Figure VII-4, Panel (e)). 2050, EUROPOP2010 assumes a net immi- The primary deficit in 2050 is projected to be gration of around 5,000 persons per year. EU- 0.75 percent under this scenario, compared to ROPOP2010 also differ from the United 0.83 percent under the EUROPOP-based Population projections with respect to fertili- LFP scenario, and about 0.91 percent under ty rate and life expectancy.93 As a result of the baseline and retirement-age scenarios. these assumptions, the population in 2050 is This translates into lower ratios of debt to somewhat different. According to EURO- GDP for the high-LFP scenario (43.7 per- POP2010, Bulgaria’s population would shrink cent), and the EUROPOP-based LFP sce- to 5.899 million by 2050, while the UN proj- nario (49.4 percent) in comparison to the ects a lower population of 5.457 million, baseline and retirement-age scenarios (51.3 largely driven by the assumption of higher em- percent and 50.8 percent, respectively). igration. Higher labor-force participation proves to be an effective way to coun- Education Scenarios teract the growth-dampening effects of Since investment in education is a key the demographic transition. The High channel through which the government LFP scenario significantly boosts GDP could boost the productivity growth, the growth (Figure VII-4, Panel (b)). On average, model presents two different education this effect is about 0.5 percent per year, which scenarios. The baseline and High LFP pro- increases the GDP per capita from about jections presented above assume that student 22.200 leva under the baseline to about enrollment rates remain constant. The high- 26.400 leva in 2050. Under the retirement- education scenario presented here, on the age scenario, GDP growth increases by about other hand, assumes a steady increase in the 0.3 percentage points per year between 2020 share of students that complete secondary or and 2025. However, once the reform is com- tertiary degrees. As a result, contrary to the pleted, the growth rate of the labor force is standard LFP scenarios, this scenario also as- determined solely by the demographic dy- sumes an increase in the share of population namics again. Therefore, GDP growth con- with tertiary and secondary education. Not- verges to baseline growth for the remaining withstanding these differences, the baseline projection period. In comparison, the EU- and the high education scenarios are based on ROPOP based LFP scenario contributes on the same United Nations population projec- average about 0.18 percent to the growth rate tions.This imposes a rather strong assumption, per year, slightly increasing towards the end of the projection period.94 Higher labor force participation im- 93 See Annex I for a detailed description of the differ- proves fiscal sustainability. With higher ence between Eurostat and United Nations popula- LFP comes higher income per capita. This tions projections. 94 The EUROPOP based LFP scenario assumes, simi- leads to a small increase in public health-care lar to the baseline scenario LFP rates by gender and expenditures as a share of GDP for the rea- five-year age groups in line with ILO projections up to sons explained in the previous scenario. How- 2020 and constant LFP rate by gender and five-year ever, the budget transfers to the pension fund age groups thereafter. 86 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Figure VII-5: Labor force by education level under different scenarios Labor Force (in Millions) 2 1.5 1 0.5 0 15–39 40–64 65+ 15–39 40–64 65+ 15–39 40–64 65+ 15–39 40–64 65+ 2010 Baseline Maximum LFP High Education Primary Secondary Tertiary Source: WB Staff simulations based on data from UN population, ILO and the Vienna Institute of Demography. Projections are for 2050. The high education scenario assumes constant LFP rates by age and education. since education is likely to affect all three fun- is bigger under the high-education scenario. damental demographic components: fertility, This provides an additional source of growth mortality and migration. The number of chil- to the increase in skills. On average, this sce- dren, the timing of births and marriage tend nario generates an additional 0.23 percent to differ among women with different levels growth rate after it becomes effective around of education. Life expectancy tends to be 2020. Later, it becomes stronger and reaches higher among the more educated. Moreover, almost 0.4 percentage points in 2050 the highly educated tend to be more likely to (Figure VII-6, Panel (a)) migrate and are less likely to return to their Higher school enrollment increases country of origin. All these suggest that pop- public education expenditures com- ulation projections could vary under different pared to the baseline. Instead of decreasing education scenarios, but the direction of the as in the baseline case, public education ex- change is ambiguous. Although the total pop- penditures increase from a little above 4.6 ulation is assumed to be the same under the percent in 2012 to near 5.1 percent in the baseline and the high education scenario, the medium term and fall back to 4.5 percent by labor force is projected to be larger under the 2050 (Figure VII-4 – Panel (c)). The increase high-education scenario since labor force in fiscal costs is mitigated by the assumption participation rates tend to be higher for the that investment in education reduces demand population with higher education within a for other public services such as police and given age group (Figure VII-5).95 hospital care. This argument is supported by Can improving the skills of the la- empirical evidence that higher education bor force compensate for its reduction tends to be correlated with lower crime rates in size? By 2050, the number of students and more frequent practice of preventive graduating with a tertiary-school diploma health-care (Lochner and Moretti 2004). reaches 204,000 under the high education scenario, which compares to 88,000 under 95 This observation is almost universal. Even among the baseline.96 This leads to an increase in the richer EU countries with significantly higher aver- the average years of schooling from about 11 age years of education than Bulgaria, labor force par- ticipation rates tend to increase with education level. years in 2012 to 12.2 years by 2050. In addi- 96 Similarly, the number of students graduating with a tion, as labor force participation is higher secondary school diploma will increase from 265 thou- among the higher-educated, the labor force sand to 440 thousand. Macro-Fiscal Implications of Bulgaria’s Demographic Change | 87 Figure VII-6: Projections under Education Scenarios a. School Enrollment b. Labor Force 600 3.6 3.4 500 Labor Force (Millions) 3.2 School Enrollment 400 (Thousands) 3.0 300 2.8 2.6 200 2.4 100 2.2 0 2.0 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 Baseline Tertiary Baseline Secondary Baseline LF High Education Scenario c. Public Expenditures for Education d. GDP Growth 5.4 3.0 Public Education Expenditures 5.0 GDP Growth Rate (%) 2.5 4.6 2.0 (% of GDP) 4.2 1.5 3.8 1.0 3.4 3.0 0.5 2.6 0.0 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 Education Baseline Higher Education Scenario Education Baseline Higher Education Scenario e. Primary Balance f. Debt to GDP Ratio 0.4% 0.55 0.2% 0.45 Debt to GDP Ratio 0.0% Primary Balance (% of GDP) –0.2% 0.35 –0.4% –0.6% 0.25 –0.8% –1.0% 0.15 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 Education Baseline Higher Education Scenario Education Baseline Higher Education Scenario Source: Authors’ simulations. 88 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria A better-educated workforce helps These simulations suggest that Bul- make public finances more sustainable. garia’s demographic change poses a seri- On average, a more educated and productive ous threat to growth and fiscal workforce leads to a reduction in the prima- sustainability in the absence of reform. As ry deficit by about 0.15 percentage points the high-LFP scenario shows, increasing LFP per year. As a result, the debt-to-GDP ratio rates could significantly raise growth and im- decreases from 51 percent in 2050 under the prove the public-debt trajectory (Table VII-3). baseline to near 41 percent under this sce- Public investment in education is also project- nario.This is to some extent driven by the as- ed to lead under the high education scenario sumption that education investment raises to significant gains in terms of growth and fis- living standards as public services other than cal sustainability if education raises productiv- education also improve as the skills of the ity, LFP rates and efficiency of public spending. work-force improve. The low-TFP scenario illustrates that down- As discussed in Chapter III, the Bul- ward risks are significant. Maintaining strong garia education system faces several productivity growth, which will entail active challenges that are not reflected in these polices and planning, is essential to maintain simulations: First, in the future, Bulgaria will debt at a sustainable level. need to decrease its high drop-out rates, espe- The debt-to-GDP ratio increases cially among the Roma and other disadvan- particularly under the low-productivi- taged groups. This is likely to require ty growth scenario through its impact additional public funding. On the other hand, on the primary deficit. Under this sce- Bulgaria would need to develop an effective nario, the primary deficit adds 1 percentage program of life-long learning. Costs of such a point per year on average to the debt-to- program are not reflected in the simulations. GDP ratio whereas it does not exceed 0.8 They are also based on the assumption that the percentage points under any other scenario costs of education per child in terms of GDP (Table VII-4). Under the higher education per capita do not increase under the high-ed- and higher labor force participation, the ucation scenario. Second, the model assumes dampening-growth effect gains importance. that school attainment in itself increases the For example, under the higher-education productivity of the workforce. Again, this as- scenario, this effect reaches 0.9 percent per sumption may not hold if the education sys- year towards the end of the projections, tem does not serve the business needs or if the nearly compensating the pressures created quality of the education decreases as the num- by primary deficits. Overall, the largest av- ber of graduates rises. erage debt-dampening effect of growth comes from higher labor force participation. Macro-Fiscal Implications of Bulgaria’s Demographic Change | 89 Table VII-3: Growth rates and public debt ratio under baseline and scenarios 2012 2015 2020 2025 2030 2035 2040 2045 2050 GDP Growth Baseline 0.8% 2.2% 1.7% 1.3% 1.2% 1.1% 1.0% 0.8% 0.7% Low TFP 0.8% 1.7% 1.5% 1.0% 1.0% 0.9% 0.7% 0.6% 0.4% High TFP 0.8% 3.0% 2.5% 1.7% 1.7% 1.6% 1.4% 1.1% 0.8% Retirement Age 0.8% 2.2% 1.7% 1.3% 1.2% 1.2% 1.0% 0.8% 0.7% High LFP 0.8% 2.2% 1.7% 1.8% 1.8% 1.8% 1.7% 1.5% 1.3% EUROP based LFP 0.8% 3.1% 1.7% 1.1% 1.2% 1.4% 1.3% 1.1% 0.9% High Education 0.8% 2.1% 1.6% 1.6% 1.5% 1.3% 1.2% 1.1% 1.0% GDP per capita Growth Baseline 1.5% 2.9% 2.5% 2.1% 2.1% 2.0% 1.8% 1.6% 1.5% Low TFP 1.5% 2.4% 2.2% 1.8% 1.9% 1.8% 1.6% 1.4% 1.3% High TFP 1.5% 3.7% 3.2% 2.6% 2.6% 2.4% 2.2% 2.0% 1.7% Retirement Age 1.5% 2.9% 2.4% 2.1% 2.1% 2.0% 1.8% 1.6% 1.6% High LFP 1.5% 2.9% 2.5% 2.6% 2.7% 2.7% 2.5% 2.3% 2.2% EUROPOP based LFP 1.5% 3.3% 2.4% 1.9% 2.0% 2.0% 1.8% 1.7% 1.5% High Education 1.5% 2.8% 2.4% 2.4% 2.3% 2.2% 2.0% 1.9% 1.9% Debt to GDP Ratio Baseline 18.0% 23.4% 29.3% 32.7% 35.9% 38.9% 41.9% 46.1% 51.4% Low TFP 18.0% 23.7% 31.5% 36.7% 41.6% 46.2% 50.9% 56.7% 63.8% High TFP 18.0% 22.7% 28.4% 31.3% 33.8% 35.7% 37.5% 40.4% 44.7% Retirement Age 18.0% 23.4% 29.2% 32.3% 35.5% 38.4% 41.4% 45.5% 50.9% High LFP 18.0% 23.4% 29.3% 32.0% 34.5% 36.4% 38.1% 40.5% 43.7% EUROPOP based LFP 18.0% 23.2% 29.1% 32.6% 35.8% 38.4% 40.8% 44.2% 48.5% High Education 17.3% 22.2% 27.3% 29.6% 31.7% 33.4% 35.2% 37.8% 41.3% Table VII-4. Comparison of Debt Dynamics across the Scenariosa Change in Debt to GDP Ratio Growth Effect Interest Payments Primary Balance Baseline 0.8% –0.4% 0.5% 0.8% Low TFP Growth Scenario 1.2% –0.4% 0.5% 1.0% High TFP Growth Scenario 0.6% –0.5% 0.4% 0.7% High LFP Scenario 0.6% –0.6% 0.4% 0.7% EUROPOP based Scenario 0.8% –0.5% 0.5% 0.8% High Education Scenario 0.6% –0.4% 0.4% 0.6% a Annual averages in percentage points, 2012–2050. 90 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Conclusion Addressing the economic impact of of aging. Providing the public with better population aging requires a combina- information about the economic conse- tion of prudent fiscal policy and provi- quences of aging and inviting an open pub- sion of high-quality public services lic debate about the cost of inaction, available with a sustained commitment of im- options, and pros and cons of different re- plementing broad-based reforms. Pro- forms might be an important step. viding high-quality public services in areas One of the most important choices such as basic education, life-long learning that the Bulgarian society will have to and health will be increasingly important as make is how to fund the needs of the Bulgaria’s population ages. Combined with elderly. Possible options, discussed in this targeted policies that activate population report, are to further increase the retirement groups with traditionally low LFP rates and age, provide public transfers to older people improvements in the business climate, these or to foster the accumulation of household policies could boost employment rates. They assets. Bulgaria is unlikely to be able to avoid would also most likely reduce emigration additional increases of the legislated retire- and stimulate the economy. ment age in the future. But for older work- There are no easy or quick solu- ers to find productive jobs, it will be essential tions. There is an urgency to start the re- to implement productivity-enhancing poli- form process. Institutions and policies take cies, such as improving the business climate, time to evolve, but the elderly of the future infrastructure and innovation, and provide are already entering the education system or them with better life-long learning oppor- work-force today. Current institutions influ- tunities, improvements in the health-care ence their choices, and thereby, the coun- system and labor-market policies targeted to try’s future growth. The slower the reform them. Moreover, strengthening the quality process, the heavier the effects of Bulgaria’s of Bulgaria’s basic education system can help demographic change are likely to become. ensure that the young people of today have Some of the reforms discussed in this report, the skills they need to be productive when such as in health-care, are likely to be polit- they become old. Given the unsustainability ically challenging to implement. Moreover, of public-transfer systems, there is an urgen- one-off reforms are unlikely to be enough. cy to convince current workers to save more The report shows that sustained reform for their future and to incentivize household commitment in all relevant areas will be re- savings. But unless economic growth accel- quired to tackle the economic consequence erates and unemployment declines, this will 91 be admittedly a daunting task. There is no retirement and increasing asset-accumula- doubt that population aging will challenge tion it should be possible for Bulgaria to public and private budgets. By investing in meet this challenge. future productivity growth, postponing 92 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria References Adams, Robert. 2002. “Social Policy For tions for a sustainable solution to the ed- Social Work”. London, UK. ucational problems of minorities.” at Atanasova, E., M. Pavlova, et al. 2011. http://www.ncedi.government.bg/Re- “What have 10 years of health insurance port percent20ASA percent202003 per- reforms brought about in Bulgaria? Re- cent20belejki_1 percent20 percent281 appraising the Health Insurance Act of percent29.htm 1998.” Health Policy 102(2–3), 263–269. Bloom, David, David Canning and Günther Baeza, Christian, and Truman Packard. Fink. 2010. “Implications of population 2006. “Beyond Survival. Protecting ageing for economic growth,” Oxford Households from Health Shocks in Lat- Review of Economic Policy, Oxford Uni- in America.” Washington, D.C.: The versity Press, vol. 26(4), pages 583–612. World Bank. Bloom, David, David Canning, Günther Basu, Alaka. 2002. “Why does Education Fink and Jocelyn Finlay, 2009. “Fertili- Lead to Lower Fertility? A Critical Re- ty, female labor force participation, and view of Some of the Possibilities”, World the demographic dividend,” Journal of Development Vol. 30, No. 10, pp. Economic Growth, Springer, vol. 14(2), 1779–1790, 2002. pages 79–101, June. Becker, S., Kevin, M., Tamura, R., 1990. Bloom, David, David Canning and B. Gra- “Human capital, fertility and economic ham. 2003. “Longevity and Life Cycle growth.” Journal of Political Economy, 98, Savings”, Scandinavian Journal of Econom- 12–37. ics, 105, 319–38. Betcherman, Gordon, Meltem Daysal and Bogdanov, Latchezar and Georgi Angelov. Carmen Pages. 2008. “Do Employment 2006. “Roma Integration in Bulgaria: Subsidies Work? Evidence from Region- Necessary Reforms and Economic Ef- ally Targeted Subsidies in Turkey”, IZA fects,” Open Society Institute Sofia, So- Discussion Paper No. 3508, Institute for fia, Bulgaria. the Study of Labor, Bonn, Germany. Campos, Nauro F. and Coricelli, F. 2002. Bulgarian International Centre for Minority “Growth in Transition: What We Know, Studies and Intercultural Relations. 2003. What We Don’t, and What We Should.”, “Evaluation of existing educational poli- Journal of Economic Literature, 40 (3), pag- cies and practices for the provision of es 793–836. equal access to education of minority Chakraborty, Sarbani, Maria Hofmarcher, children and to develop recommenda- and Johannes Koettl. 2010. “An 93 Economic Framework for Financing, European Commission. 2012a. “The 2012 Provision, and Regulation of Long- Ageing Report: Economic and budget- term Care Services for Older Popula- ary projections for the EU-27 Member tions.” Washington, D.C.: World Bank. States (2010–2060)”, European Econo- Chawla, Mukesh, Gordon Betcherman, and my 2, 2012, Brussels, Belgium. Arup Banerji. 2007. “From red to gray: http://ec.europa.eu/economy_finance/pub- the third transition of aging populations lications/european_economy/2012/pdf/ in Eastern Europe and the former Sovi- ee-2012–2_en.pdf et Union,” Washington, DC. ———. 2012b. “The 2012 Ageing Report: Couffinhal, Agnes. 2012. “Health Sector Underlying Assumptions and Projec- Reform Policy Note – Bulgaria 2012”. tions Methodologies”, European Econ- The World Bank. Mimeo. omy 4, 2011, Brussels, Belgium. http:// Cutler, D. and A. Lleras-Muney. 2006. “Ed- ec.europa.eu/economy_finance/publi- ucation and Health: Evaluating Theo- cations/european_economy/2011/pdf/ ries and Evidence”, NBER Working ee-2011–4_en.pdf Paper 12352, Cambridge, MA. ———. 2012c. “Employment and Social Danish Ministry of Education. 2007. “Den- Developments in Europe 2012,” The mark’s strategy for lifelong learning – Directorate-General for Employment, Education and lifelong skills upgrading Social Affairs and Inclusion, Brus- for all,” Report to the European Com- sels, Belgium. mission ———. 2012d. “Discrimination in the EU Daveri, F., and M. Maliranta. 2007. “Age, in 2012.” Special Eurobarometer 393, No- Seniority and Labor Costs.” Economic vember, Brussels, Belgium. Policy 22 (1): 117–75. ———. 2012e. “Europe 2020 targets: Em- Deaton, Angus. 2007. “Income, Aging, ployment rate” http://ec.europa.eu/eu- Health and Wellbeing Around the r o p e 2 0 2 0/p d f /t h e m e s /15 _ e m p l o World: Evidence from the Gallup World yment_target.pdf Poll,” NBER Working Papers 13317, _____.2012f. “Lifelong learning statistics.” Cambridge, MA. http://epp.eurostat.ec.europa.eu/statis- Dedry, A., E. Ley, H. Onder, and P. Pes- tics_explained/index.php/Lifelong_ tieau. 2013. “Aging, Migration, and learning_ statistics. Capital Accumulation”, unpublished. ———. 2012g. “New skills and jobs in Eu- Deliktas, E. and Balcilar, M. 2005. “A Com- rope: Pathways towards full employ- parative Analysis of Productivity ment.” http://ec.europa.eu/research/so Growth, Catch-up, and Convergence in cial-sciences/pdf/new-skils-and-jobs-in- Transition Economies.”, Emerging Mar- europe_en.pdf kets Finance and Trade, 41(1), pages 6–28. ———. 2011. “Estimating the capital stock Dimova A, Rohova M, Moutafova E, Atana- for the NUTS2 regions of the EU-27”, sova E, Koeva S, Panteli D, van Ginnek- European Union Regional Policy Work- en E. 2012. “Bulgaria: Health system ing Papers, No.1/2011. http://ec.europa. review.”, Health Systems in Transition, eu/regional_policy/sources/docgener/ 14(3):1–186, http://www.euro.who.int/ work/2011_01_capital_stock.pdf __data/assets/pdf_f ile/0006/169314/ ———. 2010a. Report: “Patient Safety and E96624.pdf Quality of Health Care.”. Eurobarometer. Doyle, Rachel, et al. 2009. “How can health 327(72.2), p.55 systems respond to population ageing?” ———. 2010b. “New Skills for New Jobs: World Health Organization, Policy Action Now”. A report by the Expert Brief 10. Group on New Skills for New Jobs 94 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria prepared for the European Commission. ———. 2009. Opportunities and challenges Retrieved 25 February 2013 from for labour mobility in times of crisis and http://ec.europa.eu/education/focus/fo- beyond. Retrieved 23 March 2013 from cus2043_en.htm. http://www.ilo.org/public/english/re- ———. 2009a. “2009 Ageing Report. Eco- gion/eurpro/geneva/download/events/ nomic and budgetary projections for the ministers2009/opportunities_challeng- EU-27 Member States (2008–2060)”, es_en.pdf European Economy 2, 2009, Brus- ———. 2008. “Skills for improved produc- sels, Belgium. tivity employment growth and develop- http://ec.europa.eu/economy_finance/pub- ment”. Retrieved 23 March 2013 from lications/publication14992_en.pdf http://www.ilo.org/wcmsp5/groups/ ———. 2009b. “Attitudes of Europeans To- public/@ed_norm/@relconf/documen ward Corruption – Full Report”. Euro- ts/meetingdocument/wcms_092054. barometer. 325(72.2): p. 30. pdf Eurostat. 2013. “Eurostat Statistics Data- International Monetary Fund. 2012. “Bul- base”. Available at http://epp.eurostat. garia: Article IV Consultation”, Wash- ec.europa.eu/portal/ page/portal/statis- ington, DC: IMF tics/search_database. ———. 2010. “Bulgaria: Selected Issues”, Eurofound. 2012. “NEETs – Young people Country Report No. 10/159, Washing- not in employment, education or train- ton, DC: IMF. ing: Characteristics, costs and policy re- Johnson, Richard and David Neumark. sponses in Europe,” Publications Office 1997. “Age Discrimination, Job Separa- of the EU, Luxembourg. tion, and Employment Status of Older Feyrer,J. 2008. “Aggregate Evidence on Workers: Evidence from Self-Reports,” the Link between Demograpics and Journal of Human Resources, Vol. 32, Productivity”, Population and Develop- no. 4. ment, December 2008, 34 supplement, KC S, Barakat B, Goujon A, Skirbekk V, 78–99. Sanderson WC and W Lutz. 2010. Pro- Gavrilova, Natalia and Leonid Gavrilov. jection of populations by level of educa- 2009. Rapidly Aging Populations: Rus- tional attainment, age, and sex for 120 sia/Eastern Europe. International Hand- countries for 2005–2050. Demographic book of Population Aging. Chicago, Il. Research, vol. 22 (15), pp. 383–472. Gill, I. and M. Raiser. 2012. “Golden Kotlikoff, L. and L.H. Summar. 1981. The Growth: Restoring the Lustre of the Adequacy of Savings. NBER Working European Economic Model”, World Paper 627, National Bureau of Econom- Bank, Washington, D.C. ic Research, Cambridge, MA. Gragnolati M., O. Jorgensen, R. Rocha and Koettl, Johannes and Michael, Weber. 2012. A. Fruttero. 2011. “Growing Old in an “Does Formal Work Pay? The Role of Older Brazil: Implications of Population Labor Taxation and Social Benefit De- Aging on Growth, Poverty, Public Fi- sign in the New EU Member States,” nance, and Service Delivery”, Washing- IZA Discussion Papers 6313, Institute ton, D.C. for the Study of Labor, Bonn, Germany. Health Consumer Powerhouse. 2012. “Eu- Koytcheva, Elena and Dimiter Philipov. rohealth Consumer Index”, 2012. 2008. Bulgaria. Ethnic differentials in International Labor Organization (ILO). rapidly declining fertility, Demographic 2011. Economically Active Population, Research, Vol. 19, Art.13, pp. 361–402 Estimates and Projections, 6th edition, Ge- http://www.demographic-research.org/ neva, Switzerland. http://laborsta.ilo.org volumes/vol19/13/19–13.pdf References | 95 Lance Lochner & Enrico Moretti. 2004. Economics, ed. Kenneth K. Kurihara, “The Effect of Education on Crime: Ev- 388–436. New Brunswick idence from Prison Inmates, Arrests, Mossialos, Elias and Sarah Thomson. 2002. and Self-Reports.” American Economic “Voluntary health insurance in the EU”, Review, American Economic Associa- Report prepared for the Directorate tion, vol. 94(1), pages 155–189. General for Employment and Social Af- Le Corre, Pierre-Yves. 2008. “Long Term fairs of the European Commission. Care Insurance: French and Interna- Brussels, Belgium: European Commis- tional Perspective.” Presentation given sion. at the 5th Geneva Association Health & Onder, H., Pestieau, P., Ley, E. 2013. “Mac- Ageing Conference on Long Term Care roeconomic and Fiscal Implications of - Risk Profiles, Determinants and Fi- population Aging in Bulgaria”, unpub- nancing. London, U.K. November 6 lished manuscript. and 7. Organisation for Economic Development Lee, Ronald and Andrew Mason. 2011. and Co-operation (OECD). 2012a. “In- “The Price of Maturity”, Finance & De- ternational Migration Outlook”, Par- velopment, June 2011, Vol. 48, No. 2. is, France. Levine Phillip, Olivia Mitchell and John ———. 2012b. “Looking to 2060: Long- Phillips. 1999. “Differentials Between term global growth prospects.” http:// Men and Women”, NBER Working www.oecd.org/eco/outlook/2060%20 Paper No. 7243. July 1999, Cam- policy%20paper%20FINAL.pdf bridge, MA. ———. 2008. “Conceptual Framework and Lewis, S. and F. Messy. 2012. “Financial Ed- Definition of Long-term Care expendi- ucation, Savings and Investments: An ture.” http://www.oecd.org/dataoecd/ Overview”, OECD Working Papers on 24/58/ 40760216.pdf. Finance, Insurance and Private Pen- Rangelova Rossitsa and Katia Vladimirova. sions, No. 22, OECD Publishing.http:// 2004. “Migration from central and east- dx.doi.org/10.1787/5k94gxrw760v-en ern Europe: the case of Bulgaria”, SEER Mansoor, A., Quillin, B. 2007. Migration – South-East Europe Review for Labor and Remittances: Eastern Europe and and Social Affairs, issue: 03, pp. 7–30. the former Soviet Union. Washington Republic of Bulgaria. 2013. “Strategy for D.C.: The World Bank. Prevention of Early School Leavers”, Ronald, Lee and Andrew Mason. 2010. Sofia, Bulgaria “Some macroeconomic aspects of glob- ———. 2012a. “National Strategy for De- al population aging,” Demography, vol. mographic Development in the Repub- 47(1), pages S151-S172. lic of Bulgaria 2012–2030”, Sofia, Ministry of Finance of the Republic of Bul- Bulgaria garia. 2012.“Budget Forecast for the ———. 2012b. “National Concept Paper for 2013–2015 Period”, Sofia, Bulgaria. the Encouragement of Active Life of the Mitra, Pritha and Cyril Pouvelle. 2012. Elderly”, Sofia, Bulgaria “Productivity Growth and Structural ———. 2012c. “National Program for Ac- Reform in Bulgaria: Restarting the tive Life of the Elderly”, Sofia, Bulgaria Convergence Engine,” IMF Working ———. 2011. “National Reform Pro- Paper, WP/12/131, Washington, DC. gramme 2011–2015”, Sofia, Bulgaria. Modigliani, F. and R.H. Brumber. 1954. ———. 2010. “National Youth Strategy “Utility Analysis and the Consumption 2010–2020.”, Sofia, Bulgaria. Function: An Interpretation of the Sanigest Europe. 2008. “Advising on Op- Cross-Section Data”. In Post-Keynesian tions for Improving GP Referral and 96 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Hospital Admissions Practices. Analysis Development Unit, Europe and Central of referral and admission practices.} Asia Region, Washington, DC. Skirbekk, V. 2004. “Aging and individual ———. forthcoming.” Workforce Develop- productivity: A literature survey.” In: G: ment in Bulgaria: Systems Approach for Feichting (Ed.) Vienna Yearbook of Better Education Results – Workforce Population Research. Vienna, Verlage Development (SABER-Wf D) Country der Oesterreichischen Akademie der Report.”, Washington, DC. Wissenschaften ———. 2013. “The Inverted Pyramid: Pen- Smith P, Mossialos E, Leatherman S and Pa- sion Systems in Europe and Central Asia panicolas I. 2009. “Performance Mea- Facing Demographic Challenges,” surement for Health System Washington, DC. Improvement: Experiences, Challenges ———. 2012a. “Doing Business 2013. and Prospects”, Cambridge University Smarter Regulations for Small and Me- Press, London dium-Size Enterprises.”, Washing- Solow, R. M. 1956. “A Contribution to the ton, DC. Theory of Economic Growth”, Quar- ———. 2012b. “Going for smart growth: terly Journal of Economics, 70 (1), p. making research and innovation work in 65–94. Bulgaria, Washington”, DC. The European Parliament, Directorate- ———. 2012c. “Bulgaria: Public Expendi- General for Internal Policies. 2012. “La- ture Review for Growth and Competi- bour Market Situation and Pension tiveness”, Washington”, DC. System in Bulgaria”. At http://www.eu- ———. 2012d. “Bulgaria: Household Wel- roparl.europa.eu/delegations/en/stud- fare during the 2010 Recession and Re- iesdownload.html?languageDocument covery.”, Washington, DC. =EN&file=76451 ———. 2012e. “World Development Re- UNDP/WB/EC. 2011. “Regional Surveys port 2013: Jobs”, Washington, DC. among the Roma Population.” ———. 2012f. “Women, Business and the United Nations, Department of Economic Law 2012. Removing Barriers to Eco- and Social Affairs, Population Division. nomic Inclusion.”, Washington, DC. 2011. “World Population Prospects: The ———. 2012g. “Strengthening Higher Edu- 2010 Revision.” New York http://esa. cation in Bulgaria”, Washington, DC. un.org/unpd/wpp/index.htm http://siteresources.worldbank.org/ Vassilev, Rossen. 2005. “Bulgaria’s Demo- B U L GAR I AE X T N / R e s o u r c - graphic Crisis: Underlying Causes and es/305438–1307440973243/Bulgaria- Some Short-term Implications”, South- Higher-Education-executive-summary- east European Politics, Vol. VI, No. 1 March-2012-EN.pdf Więckowska, Barbara. 2006. “Dobrowolne ———. 2012e. “Toward an Equal Start: ubezpieczenia pielęgnacyjne (niedołęstwa Closing the Early Learning Gap for starczego) – niedaleka przyszłość czy pol- Roma Children In Eastern Europe”, ska science fiction?”, Prawo Asekura- Washington, DC. http://go.worldbank. cyjne, Grudzień. org/F2QMFQ7FJ0. World Bank. forthcoming. “From Jobless ———. 2011a.“The Jobs Crisis: Household Growth to Growing Jobs: Fostering Em- and Government Responses to the Great ployment Creation in Eastern Europe Recession in Eastern Europe and Cen- and Central Asia,” Washington, DC. tral Asia”, Washington, DC. ———. forthcoming. “Getting Better: Im- ———. 2011b.“World Development Report proving Health System Results in Eastern 2012: Gender Equality and Develop- Europe and Central Asia”, Human ment,” Washington, DC. References | 97 ———. 2010.“Assessing the Quality of Edu- Union. Washington, DC. http://sitere- cation in Bulgaria using PISA 2009”, sources.worldbank.org/INTECA/Re- Retrieved 12 April 2013 from https:// s ou rce s/2 5789 6 –11678563 8950 5/ openknowledge.worldbank.org/han- Migration_FullReport.pdf dle/10986/12767 World Economic Forum. 2012. “Global ———. 2007. Migration and Remittances – Gender Gap Report 2012”, Geneva. Eastern Europe and the Former Soviet 98 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria Annex I: Population projections for Bulgaria Population projections in this report are tak- With respect to migration Eurostat assumes en from either Eurostat (EUROPOP) or the an increase in net emigration from 10,000 medium variant of the United Nations Pop- per year to 14,000 in 2020, declining steadi- ulation Division. Their projections differ ly towards the beginning of the 2030s. Be- with respect to assumptions regarding fertil- tween 2030 and 2050, Eurostat assumes net ity rate, life expectancy and migration. Eu- immigration of around 5000 per year, while rostat (UN) projects that Bulgaria TFR the UN extrapolates current immigration increases from 1.56 in 2010 to 1.65 (1.93) by trends, and assumes 10000 net emigrants per 2050. Note that the TFR in the EU27 is year throughout the projection period. currently at a low 1.7. Eurostat (UN) as- As a result of these assumptions, popula- sumes that male life expectancy in Bulgaria tion in 2050 is somewhat different. Accord- increases from 70.3 in 2010 to 82.3 in 2050 ing to Eurostat, Bulgaria’s population would and female life expectancy from 77.5 in 2010 shrink to 5.899 million by 2050, while the to 85.0 in 2050. By comparison, it took the UN projects a lower population of 5.457 average European country 45 years to in- million, largely driven by the assumption of crease male life expectancy from 70 to 80. higher emigration. Also the age structure Eurostat versus un data 6,000 5,000 4,000 3,000 2,000 1,000 0 < 15 15–64 65+ < 15 15–64 65+ < 15 15–64 65+ 2010 2030 2050 Eurostat UN Source:United Nations Population Division (2011) and Eurostat (2012). 99 differs slightly: although the working-age and Eurostat projections), the share of people population as a share of total population is aged 65 and above is higher for the EU pro- the same (56 percent in 2050 for both UN jections. 100 | Mitigating the Economic Impact of an Aging Population: Options for Bulgaria