SERBIA Social Protection Situational Analysis This project is funded by the European Union SERBIA Social Protection Situational Analysis © 2022 International Bank for This work is a product of the staff of The World Bank with external contri- Reconstruction and Development / butions. The findings, interpretations, and conclusions expressed in this The World Bank work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. 1818 H Street NW, Washington DC 20433 The World Bank does not guarantee the accuracy of the data included in Telephone: 202-473-1000 this work. The boundaries, colors, denominations, and other information Internet: www.worldbank.org shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. ii SERBIA Social Protection Situational Analysis SERBIA Social Protection Situational Analysis Contents: LIST OF FIGURES iv LIST OF TABLES iv LIST OF BOXES iv ACKNOWLEDGEMENTS v LIST OF ACRONYMS vi 1. Introduction 1 2. Poverty and labor market outcomes 3 3. Overview of Serbia’s social protection system 6 4. Social assistance to support the poor and vulnerable, persons with disabilities and families with children 8 5. Social services 12 6. Employment and labor market policies 15 7. Pension system 18 8. Lessons learned from Covid-19 20 9. Summary and reform priorities 22 LIST OF REFERENCES 25 Annex 1 27 Annex 2 28 iii SERBIA Social Protection Situational Analysis List of Figures: Figure 1: Poverty headcount, 2008-2019 3 Figure 2: Labor market indicators, Serbia and Western Balkans, 5 2018-2019 Figure 3: Access to and quality of jobs, 2010-2018 5 Figure 4: Expenditures on social protection, 2012-2019 7 Figure 5: Impact of child and social protection reforms on social 10 assistance expenditures, 2013-2020 Figure 6: Share of passive and active labor market programs, 2019 15 Figure 7: Revenues and expenditures, 2012-2020 19 List of Tables: Table 1: Overview of Serbia’s social protection system 6 Table 2: Social assistance programs in Serbia 8 Table 3: Social services by category and delivery agency 12 Table 4: Policy recommendations and knowledge gaps 23 List of Boxes: Box 1: What is a social protection system? 1 iv SERBIA Social Protection Situational Analysis Acknowledgments This report was prepared by Sara Johansson de Silva (World Bank Consultant), Jamele Rigolini (Program Leader and Lead Economist), Marijana Jasarevic (Social Protection Specialist), Gonzalo Reyes Hartley (Senior Social Protection Specialist) and Gordana Matkovic (World Bank Consultant). Sarah Coll-Black (Senior Economist) and Stefanie Brodmann (Senior Economist) also contributed to the report. The team is grateful for the outstanding administrative support provided by Bisera Nurkovic (World Bank Program Assistant). Special thanks to Darcy Gallucio for editing and providing detailed comments, as well as Elizaveta Tarasova for the cover page and graphic design. The World Bank team would like to thank the representatives of the Ministry of Labor, Employment, Veterans and Social Affairs, the Ministry of Family Care and Demography, the National Employment Service, the Pension Insurance Fund, the European Union Delegation in Serbia and UNICEF for the inputs provided and the excellent collaboration. Useful comments were provided at virtual consultations held with the govern-ment and development partners in March 2022. The analysis was made possible due to the generous funding from Europe 2020 Trust Fund by DG NEAR. The note was prepared under the guidance of Linda van Gelder (Country Director for the Western Balkans), Nicola Pontara (Country Manager for Serbia) and Cem Mete (Practice Manager, Social Protection and Jobs Global Practice, Europe and Central Asia Region). v SERBIA Social Protection Situational Analysis List of acronyms ALMP Active Labor Market Program CA Child Allowance CSW Center for Social Work ECA Eastern and Central Europe and Central Asia ESPROSS European System of integrated Social PROtection Statis- tics EU European Union FSA Financial Social Assistance GDP Gross Domestic Product ILO International Labor Organization LMIS Labor Market Information System LSG Local Self-Government LTC Long-Term Care MoESTD Ministry of Education, Science and Technology Develop- ment MoLEVSA Ministry of Labor, Employment, Veteran and Social Affairs NEET Not in Employment, Education or Training NES National Employment Service OECD Organisation for Economic Co-operation and Develop- ment PA Parental Allowance PAYG Pay-As-You-Go PIO Pensions and Disability Insurance Fund of Serbia PISA Program for International Student Assessment RSD Serbian Dinars SCR Social Card Registry SEE Southeast Europe SIPRU Social Inclusion and Poverty Reduction Unit SORS Statistical Office of the Republic of Serbia SPEED Social Protection Expenditure and Evaluation Database SPIS Social Protection Information System TIMSS Trends in International Mathematics and Science Study UN United Nations UNICEF United Nations Children's Fund WDI World Development Indicators WHO World Health Organization vi SERBIA Social Protection Situational Analysis Introduction A stronger social protection system is needed force, train workers for new forms of employment, to increase the Serbian population’s resilience to and systematically facilitate employment. A strong shocks, foster human capital growth and address social protection system builds the human capital inequities. In the past 13 years, Serbia has been to protect against economic, environmental, and hit by a global financial crisis (2008), an extreme health shocks, raising wellbeing across the lifecycle weather disaster (2014) and a global pandemic of an individual and across generations. (2020). Whether economic-, environmental- or health-derived, each of these shocks has plunged The objective of this situational analysis is to the economy into recession and disproportion- provide a brief assessment of Serbia’s social protec- ately impacted the poor and vulnerable. In addition, tion system. The exercise forms part of a regional Serbia faces mounting long-term challenges to initiative, supported by the European Union, to prosperity. Economic and social indicators remain assess social protection systems across the Western significantly below European Union (EU) averages; Balkans. The situational analysis identifies how in 2019, nearly one quarter of the population was Serbia’s social protection system performs and poor or vulnerable to falling into poverty. Low birth outlines areas for reform. In this note, social protec- rates and outmigration have resulted in a rapidly tion refers to (i) social assistance,1 (ii) social services, aging and shrinking population, a trend that will (iii) social insurance (pensions),2 and (iv) employment intensify in the future. To counter this, Serbia and labor market programs (see Box 1). The analysis will need to make major investments in human is based on existing evidence and identifies knowl- capital to increase the productivity of its work- edge gaps to be covered in further analysis. Box 1: What is a social protection system? In this note, a social protection system refers to social assistance, social insurance (pensions), social services and employment and labor market programs. Social assistance targets the poor and vulnerable. These programs reduce chronic poverty and protect people from falling into poverty, provide support for other vulnerabilities and provide support at certain points in the lifecycle. Assistance includes non-contributory programs such as cash transfers, in-kind support or other forms of targeted consumption assistance as well as family- and child-related cash benefits. Social insurance targets the population at large. These programs are intended to help people smooth income across the lifecycle and protect them from shocks. Social insurance programs are contributory and provide benefits to those that contributed to the program, with the level of benefit related to their contri- bution. Examples include old-age pensions or disability pensions. Social services target individuals and families to improve or preserve their quality of life, eliminate and mitigate the risk of adverse life circumstances and create opportunities for independent living. Foster care or elderly care are examples of these services. Employment and labor market programs improve the inclusiveness and effectiveness of labor markets. They aim to increase and/or improve the quality of labor supply, labor demand and job matching, and protect income against labor market shocks. They include job-search assistance, skills training, wage subsidies, public works and unemployment benefits. Back to table of contents 1 1 In Serbia, social assistance programs are generally understood as last-resort programs. In this report, and as per World Bank typology, they refer to all non-contributory forms of assistance. 2 Health insurance is not considered here, although it forms part of social insurance mechanisms. SERBIA Social Protection Situational Analysis The challenges posed by aging, low produc- components in key categorical programs.4 The tivity and inequality of opportunities will require government can improve service delivery, and thus systemic social protection reforms. The analysis improve access, by coordinating across institutions shows that Serbia operates a comprehensive social and between the central and local level. In parallel, protection system; however, policy priorities and Serbia needs to strengthen its public sector capacity inefficiencies in program execution reduce efficiency and mechanisms for targeting, case management, and effectiveness. Social assistance programs do and monitoring and evaluation as well as continue not sufficiently reach the poor, and the support developing information systems that support these provided does not adequately cover beneficiaries’ activities. The government needs to identify the essential needs. Emphasis on achieving pro-natal most cost-effective policy package for its pro-natal objectives through social assistance programs is objectives and to restructure programs with esca- crowding out poverty-targeted programs; within lating costs. At the same time, the government poverty-targeted programs, overly strict eligibility needs to continue revisiting the pension system criteria and case management challenges further parameters and foster retirement savings that reduce coverage. The provision of social services complement pay-as-you-go (PAYG) pensions. is still limited and varies across the country, with Finally, it is essential to build systemic resilience by many Local Self-Governments (LSGs) not providing further developing the diagnostic tools and estab- key services. Case management and referral proce- lishing the legal frameworks necessary to facilitate dures are underdeveloped and there is not sufficient immediate and adequate response to shocks. coordination across social assistance and social services to ensure client-oriented, tailored services The remainder of the note is organized as follows. or efficient resource management. Demographic The second section provides an overview of changes, together with an increase in informal poverty and labor market outcomes in Serbia. The activity, are also sharpening the trade-off between third section provides an overview of the social the coverage and adequacy of pensions and finan- protection system in aggregate, followed by more cial sustainability. Alternative and voluntary retire- detailed analysis of social assistance, social services, ment savings forms to stimulate savings among the pensions, and employment programs. A specific better-off remain underdeveloped, whereas there section is dedicated to reviewing the social protec- are no targeted measures to protect the elderly tion system’s response to the COVID-19 pandemic. poor. Employment and labor market programs The final section provides an overview of reform have very limited outreach, especially considering priorities and knowledge gaps. employment challenges for vulnerable groups. The COVID-19 pandemic also shows that the social protection system is not sufficiently shock-respon- sive. Neither information systems nor legal frame- works were in place to facilitate an upscaling of the social assistance system, by increasing coverage and adequacy of benefits to poor and vulnerable. The government implemented two one-off cash transfers (to adults and pensioners) which were universal and as such costly, but did not target poor households, poor elderly, or poor children. Within the National Employment Service (NES), active labor market programs were not reorganized to provide online skills training or address new skill gaps.3 Reforms will be needed to increase the coverage and adequacy of social protection while strength- ening its financial sustainability. Reforms need to focus on effectiveness. The government can improve the coverage and adequacy of social assistance programs by shifting spending back to poverty-tar- geted programs and/or introducing means-tested Back to table of contents 2 3 Information in this section primarily relate to the first half of 2020 during the so-called emergency situation. 4 Primarily the parental allowance. SERBIA Social Protection Situational Analysis Poverty and labor market outcomes Serbia faces a challenge making its labor markets more inclusive while increasing labor productivity. Nearly one quarter of the Serbian population is living in poverty or at risk of poverty. Poor children are over- represented, and access to health and education regresses along socio-economic lines. Specific groups (women, youth, low skill, Roma) face additional obstacles in accessing job opportunities even though access to employment is overall high, especially by Western Balkan standards. The government protected employment during the first year of the COVID-19 pandemic. Yet, the quality of jobs is too low, however, as evidenced by high informality and low labor productivity. Poverty and vulnerability to poverty remain Serbia compares favorably with Albania but has important concerns in Serbia. In 2019, nearly one higher levels of poverty than North. Macedonia quarter of the population (23.2 percent) were poor or Montenegro (Figure 1, b). World Bank estimates or vulnerable to falling into poverty.5 This repre- of the impact of the COVID-19 pandemic suggest sented an improvement by 10 percentage points that, at least over the shorter term, Serbia has likely compared to the post-financial crisis years in 2013- seen the smallest impact on consumption poverty 2014. Some 7 percent still lived in absolute poverty, among Western Balkan countries, an increase of 0.1 meaning that they were unable to cover their most percentage points, compared to 4-5 percentage basic needs.6 The share of this group of extremely points for other Western Balkan countries.7 None- poor barely changed between 2015 and 2019, theless, the COVID-19 pandemic has affected although economic conditions improved substan- vulnerable groups more than others. tially (Figure 1, a). In the Western Balkan context, Figure 1: Poverty headcount, 2008-2019 a. National poverty line: Population living in poverty, b. International poverty line: Western Balkans 2008-2019 population living below international USD 5.5 per day, 2017-2019 14.0 40 35 33.8 12.0 30 % of total population 10.0 % population 25 8.0 20 19.3 17.9 6.0 15.6 15 4.0 10 2.0 5 0.0 0 2008 2010 2018 2009 2014 2013 2017 2011 2015 2012 2016 2019 ALB MKD MNE SRB 2017 2018 2016 2019 Serbia Urban Non-urban Source: Estimates based on data from (a) Statistical Office of the Republic of Serbia (2020) and (b) World Development Indicators (2021) Note: Refers to population living below USD 5.5 international per day. Back to table of contents 3 5 The at-risk-of-poverty rate represents the share of persons whose equivalized disposable income is below the relative poverty line (60 percent of the median national equivalized disposable income). Data are from the Statistical Office of the Republic of Serbia (2019). 6 Living below RSD 12,495 (approximately USD 115). Refers to consumption expenditures per month. 7 World Bank (2021b). SERBIA Social Protection Situational Analysis In Serbia, children are more likely to live in poverty percent between 2015 and 2019. Between 2014 than any other age group, but the elderly make and 2019, employment for the population 15-64 up an increasing share of the poor.8 The share of years old grew by more than 450,000 jobs on a net children living in absolute poverty is higher than basis, and the unemployment rate declined from the national average: 8.2 percent for children 0-13 close to 20 percent in 2014 to below 11 percent in years old and 8.5 percent for children 14-18 years 2019. These jobs benefited groups that usually face old, compared to 7 percent for all age groups (2019 obstacles in accessing employment, such as women, numbers). Thus, children under 18 remain over-rep- older workers, and youth. In 2019, Serbia’s employ- resented among the poor, making up 20 percent of ment rate of 60.9 percent was at par with those of the absolute poor against 16 percent of the total Albania and above other Western Balkan countries population. With the aging of the population, the (Figure 2, a). In fact, Serbia outperforms most of its elderly make up a growing share of the absolute neighbors on key indicators for labor market access, poor, albeit increasing from 21 to 30 percent with lower inactivity rates, lower levels of unemploy- between 2011 and 2019. ment and a lower share of youth Not in Employ- ment, Education or Training (NEET) (Figure 2, b, c, Across Serbia’s regions and population groups, d). However, Serbia still lags EU countries, where health and education access and outcomes differ employment rates reached 68.5 percent in 2019.14 significantly. Although health outcomes for children During 2020, Serbia’s policy package successfully are strong overall, regions with higher poverty levels protected employment levels against the impact of (largely in southern and eastern Serbia) are more the pandemic. Employment rates even increased, likely to show worse health outcomes for children. to 61.3 percent, whereas unemployment rates fell For example, infant mortality rates and the share below 10 percent. Inactivity increased only margin- of low birthweight babies are many times higher in ally, from 31.9 to 32.4 percent of the population. poor regions.9 Despite progress on infant and child Informal employment fell by some 10 percent, mortality rates of children in Roma settlements, however, and earnings are likely to have been they remain almost twice as high as for Serbia as a affected, especially in the informal services sector. whole, and the incidence of stunting, which is asso- ciated with poorer schooling outcomes, reduced Women, youth, low-skill, and Roma face addi- productive capacity and poor health,10 is three times tional obstacles in accessing jobs. Serbian women as high.11 The attendance rate in early childhood have the highest employment rates in the Western education has increased but, at 61 percent (ages 36 Balkans region, but the gap with Serbian men has to 59 months), remains far short of the EU’s goal remained constant at around 13-14 percentage of 95 percent by 2030. Only 11 percent of children points since 2010. This gap largely reflects norms from the poorest quintile and only 7 percent of around family responsibilities: among those children living in Roma settlements are enrolled in reporting to be inactive due to care responsibilities preschool. The average child in Serbia can expect to (for children, the elderly or persons with disabilities), remain in school 13.3 years, the same length of time 93 percent are women.15 Those with low levels of as the average child in Germany. On average, Serbia education (primary school at the most) are much also performs relatively well in international learning less able to access jobs, and also more likely to hold assessments although significant variations and precarious jobs in the informal sector. One-in-five poor learning outcomes strongly correlate with young people 15-29 years old is out of school but socio-economic status and geographic location. In not enrolled in in any education or training (NEET). the 2018 Program for International Student Assess- Persons living in households where the household ment (PISA) survey, nearly four-in-ten 15-year-old head is unemployed or inactive are much more likely students were not performing even at the basic to be living in poverty than those in some form of level of reading, mathematics and science literacy.12 employment or receiving a pension.16 Estimates Weak learning outcomes affect lifelong employ- suggest that closing employment and earnings ment opportunities and living standards. gaps between Roma and the general population in Serbia could bring productivity gains in the order Access to employment is relatively high compared of between 0.9 and 3.5 percent of gross domestic with other Western Balkan countries and jobs have product (GDP).17 Similarly, closing the gender gap so far been insulated from the effects of COVID- by bringing women’s access to employment up to 19.13 Following a dip after the global financial crisis, the current level of men’s could increase per capita economic growth turned positive, averaging 3.2 gross income by as much as 16 percent.18 Back to table of contents 4 8 Poverty analysis based on estimates on data from SORS (2020). 14 Estimates based on data from Eurostat for 2019. Population 15-64 years old. 9 UNICEF and World Bank (2021a). 15, Estimates based on data from SORS (2020). 10 WHO (2014). 16 Estimates based on data from SORS (2020). 11 UNICEF (2020). 17 Robayo-Abril and Millán (2019). 12 OECD (2019). 18 Cuberes and Teignier (2016). 13 Employment analysis based on data from Southeast Europe Jobs Gateway (up to 2018) and SORS (2019, 2020). SERBIA Social Protection Situational Analysis Figure 2: Labor market indicators, Serbia and Western Balkans, 2018-2019 a. Employment rate b. Unemployment rate 80% 50% 70% 60% 40% 50% 30% 40% 30% 20% 20% 10% 10% 0% 0% WB5 ALB BIH MNE MKD SRB WB5 ALB BIH MNE MKD SRB Total 15-64 Male 15-64 Female 15-64 Total 15-64 Male 15-64 Female 15-64 Total 15-24 Male 15-24 Female 15-24 Total 15-24 Male 15-24 Female 15-24 c. Inactivity rate d. NEET rate, ages 15-24 100% 30% 80% 25% 20% 60% 15% 40% 10% 20% 5% 0% 0% WB5 ALB BIH MNE MKD SRB WB5 ALB BIH MNE MKD SRB Total 15-64 Male 15-64 Female 15-64 Total Male Female Total 15-24 Male 15-24 Female 15-24 Source: Estimates based on Southeast Europe Jobs Gateway (2019). Note: Data for employment, unemployment and inactivity are for 2019 Q2; data for NEET rate are for 2018. Labor productivity remains low, undermining the welfare gains from job creation. Productivity growth is essential to sustain improvements in earnings and living standards. However, job creation in the period from 2014 to 2019 also brought an increase in the share of precarious jobs in Serbia: informal jobs and temporary employment. Labor productivity levels have diverged from the EU in this period, with GDP per worker reaching only 45 percent of the EU average, proving that lower skill tasks and jobs dominate in the Serbian economy (Figure 3). Figure 3: Access to and quality of jobs, 2010-2018 Access to jobs Quality of jobs 2900 30 25 60 Unemployed as % of active population 58 Labor productivity, % of EU average 2800 25 20 56 Share of total employment 2700 Thousands employed 54 20 2600 15 52 2500 15 50 2400 10 48 10 46 2300 5 44 Employment 5 Temporary employees 2200 42 Unemployment rate Informally employed Labor productivity, % of EU average 2100 0 0 40 2010 2011 2012 2013 2014 2015 2016 2017 2018 2010 2011 2012 2013 2014 2015 2016 2017 2018 Source: Estimates based on WIIW/World Bank (2020) and World Bank (2021a). Note: Labor productivity is expressed in terms of GDP per worker. Back to table of contents 5 SERBIA Social Protection Situational Analysis Overview of Serbia’s social protection system The social protection system in Serbia consists of social assistance, social insurance, social services and employment and labor market programs. Expenditures on social protection are relatively high but largely captured by pensions, leaving limited resources for non-contributory social assistance, in particular pover- ty-targeted programs, social services, and employment and labor policies. The poor and vulnerable are insufficiently covered by social assistance benefits, due to the diversion of social assistance programs away from poverty-targeted and toward categorical programs, limited outreach and strict eligibility criteria that in turn reduce access to poverty-targeted programs. Investment in programs to improve labor market outcomes and labor market vulnerability is too low to make a significant impact. An overview of Serbia’s social protection system is provided in Table 1. The first section consists of social assistance programs: cash and in-kind benefits to support the poor and vulnerable, and social services. Cash and in-kind benefits include both means tested (targeted at the poor) and categorical (targeted to various characteristics). The second section presents social insurance programs. The third section includes employment programs.19 Table 1: Overview of Serbia’s social protection system Category Benefits and programs Means tested: • Financial social assistance (FSA) • Child allowance (CA) Benefits to support • One-off assistance (includes in-kind) the poor, the • Energy benefit vulnerable, persons Categorical: with disabilities • Parental allowance (PA), which is also known as a birth grant and families with • Wage compensation for maternal leave, childcare and extended childcare leave, and Social assistance children other benefits based on birth, childcare and special childcare • Preschool subsidy • War Veteran Benefit programs (includes in-kind) • Caregiver Allowance • Accommodation and shelter services: residential, foster care, shelters • Services for independent living of persons with disabilities Social services • Counseling, therapy and socio-educational services • Home care services for persons with disabilities Contributory • Contributory pension system • Farmers Social assistance Private contributory and social insurance Pensions • Voluntary private programs • Occupational pension funds Non-contributory • Cash benefit for parents who do not have a pension and who have cared for child with disability • Unemployment benefit (passive labor market program) • Active labor market programs (ALMPs) • Job-search assistance and mediation services • Professional orientation and career planning counseling Employment and labor programs • Skills development training • Wage subsidy • Support to self-employment • Incentives for FSA beneficiaries • Public works Back to table of contents 6 The main legal pillars for the social protection system are summarized in Annex 1. 19 SERBIA Social Protection Situational Analysis Social protection benefits and services are deliv- Public expenditures on social protection repre- ered by both central and local government insti- sented less than 6 percent of GDP in 2020. tutions. Some social assistance benefits and social services, for example, are delivered by local self-gov- Total social protection expenditures have fallen ernments (LSGs). Others fall under a central govern- in recent years, but the trend is driven by pension ment mandate and are delivered through Centers expenditures and is not likely to continue. In the for Social Work (CSW) and/or social welfare depart- past few years, Serbia has registered a substantial ments and/or other units in the local self-govern- decline in pension outlays, reflecting labor market ments. Some services, including for independent improvements and a temporary reform measure living, housing and shelters, are delivered by LSGs (imposed between 2014 and 2018) to contain and CSWs, while other services are outsourced pension expenditures. These trends are not likely to non-governmental organizations (NGOs) and to be sustained in the future as the demographic private providers. decline continues and upward pressure on pension expenditures will ensue. Expenditures on social protection are relatively high but are dominated by high spending on Social protection expenditures are less pover- pensions. Using 2019 (pre-Covid) data, Serbia’s ty-targeted than in many European peer countries. expenditures on social protection are similar to This is partly driven by the high share of contribu- those of new EU member states and exceed those tory social insurance expenditures benefitting those of other Western Balkan countries, save Monte- with long and continuous employment history. negro (Figure 4). However, most social protection However, it also reflects a limited poverty focus of spending is absorbed by social insurance (10.3 social assistance programs. Serbia spends 5 percent percent of GDP in 2020, or 71 percent of total social of social protection expenditures on means-tested protection spending), leaving limited resources for (poverty-targeted) programs. The share is high social assistance and employment and labor market compared to Western Balkans (for example, around programs.20 Social insurance contributions (from 2 percent in Montenegro and North Macedonia) employees and employers) make up the major but low compared to EU countries for which the share of social protection expenditures (8.7 percent average share of expenditures reaches 12 percent of GDP, or more than 60 percent of expenditures). (Figure 4).21 Figure 4: Expenditures on social protection, 2012-2019 16 14 12 10 % of GDP 8 6 4 2 0 SRB 2019 BIH 2017 SLO 2017 LIT 2016 ALB 2018 CZE 2017 LAT 2017 EST 2016 HUN 2016 MKD 2017 CRO 2014 BUL 2017 POL 2012 ROM 2017 MNE 2012 Social insurance Social assistance Labor market programs Social services Source: Estimates based on Social Protection Expenditure Evaluation and Database (2022). Back to table of contents 7 20 Recent budget expenditures on social services are not available. 21 Estimates based on European System of Integrated Social PROtection Statistics (ESPROSS) data (2022). SERBIA Social Protection Situational Analysis Social assistance to support the poor and vulnerable, persons with disabilities and families with children Serbia needs to redress the imbalances in programmatic expenditures and increase the effectiveness of poverty-targeted programs to reduce poverty and vulnerability. Coverage of the poor has improved but more than half of the population in the poorest income quintile still do not receive any benefits. Despite a policy focused on assisting families with children, half of the children under age 18 who live in extremely poor households are not covered by social assistance benefits at the household level. Indeed, social assis- tance skews heavily toward pro-natal benefits that are not poverty targeted. Poverty-targeted benefits are also low compared to the needs of poor families. The COVID-19 shock showed that the social assis- tance system lacks the capacity to scale up to help households weather shocks. The current programmatic landscape crowds out effective programs and threatens the sustainability of social assistance financing. Serbia has a comprehensive system of social assis- local self-governments can design these benefits is tance programs targeted at many different groups uneven across areas. and delivered and financed by different institu- tions. Programs are directed at, inter alia, low-in- The main social assistance programs fall into two come households, households with low access to categories: means tested and categorical (Table 2). affordable energy, persons with disabilities, war The main means-tested programs are: (i) financial veterans, working mothers, and families with chil- social assistance (FSA), (ii) child allowance (CA), and dren. The Centers for Social Work (CSW) and social (iii) energy benefit. There are numerous categori- welfare departments and/or other units in local cal programs: wage compensation during mater- self-governments (LSGs), among other, are involved nity leave, childcare and extended childcare leave, in the targeting and delivery of programs. Most parental allowance (PA), caregiver allowance and cash benefits that are subject to national regula- war veteran benefits. Preschool subsidies combine tion are financed from the national budget. The categorical and means-tested eligibility. Table 2: Social assistance programs in Serbia Poverty-Targeted Programs Non-Poverty-Targeted Programs (Means Tested) (Categorical) Financial social assistance (FSA) provides income Wage compensation during maternity, childcare and support to families or households that meet the extended childcare leave, and other benefits based eligibility criteria related to income, asset ownership on the birth of a child, childcare and special childcare and unemployment status of able-bodied members. compensate mothers who are formal employees (temporary or permanent) for lost earnings during Child allowance (CA) supports the income of maternity leave. As of 2018, women who had flexible poor households with children. CA is universal for work arrangements, were self-employed or were paying house-holds who have children with disabilities and/or into a farmer pension became eligible for a similar set of children who receive the Caregiver Allowance. CA benefits. Benefits are proportional to wages/earnings. means tests eligibility for poor households who have children. If the household receives FSA, the children Parental allowance (PA) consists of a lump-sum are required to attend school payment for first-born children; 24 monthly install- ments for second-born children; and as of 2018, 120 monthly installments for third- and fourth-born children. In 2018, the government also introduced immunization, preschool requirements and primary school requirements. Back to table of contents 8 SERBIA Social Protection Situational Analysis One-off assistance is financed and administered Caregiver allowance is for people who, due to physical locally; it is not uniform across the country. In order to or sensory impairment, intellectual disability or changes be eligible, families usually have to provide documents in health need the help and care of another person in “proving the situation of unexpected and temporary order to satisfy their basic life needs. hardship”. The benefit is provided once or twice per year in cash or in-kind and eligibility can be categorical War veteran benefits comprise income and in-kind or means tested. support for war veterans, survivors, civilian victims of war and their families. Benefits include wage compen- Energy benefit is a mean-tested program that aims sation for working veterans, cash compensation for to protect the poor by deduction of their monthly disabled veterans, caregiver allowance and survivor electricity and gas consumption bills. Local Self-gov- benefits. ernments review and determine the eligibility for the energy benefit, which creates fragmentation in benefit Preschool subsidy is provided to children with disabil- delivery. ities, children without parental care and kindergarten children whose parents are FSA beneficiaries. (This subsidy can be considered a hybrid between means tested and categorical.) Public spending on social assistance has fallen and insurance programs that are financed through con- its composition has shifted further away from pov- tributions. With spending of RSD 26 billion in 2019, erty-targeted programs. In 2019 and 2020, social wage compensation during maternity, childcare assistance spending reached respectively 1.83 and and extended childcare leave represented 26 1.96 percent of GDP, similar to many EU countries percent of total social assistance expenditure— with significantly lower poverty rates. Expenditures twice the amount spent on FSA. on social assistance have steadily fallen since 2013 when they amounted to 2.07 percent of GDP. These Based on current trends, the expansion of pro-na- trends are driven by a significant reduction in spend- tal benefits may threaten the sustainability of ing on poverty-targeted expenditures, from 0.63 to social assistance system financing. Projections for 0.47 percent of GDP in 2013 and 2019 respectively, the PA indicate that even with unchanged fertility and an increase in spending on categorical pro- rates, the projected budget may surpass the current grams, from 1.43 to 1.52 percent of GDP in the same budget for all social assistance programs in 10 time period. Categorical programs now account years. Meanwhile, expenditures on the CA program for 77 percent of total social assistance expendi- have slightly fallen whereas FSA expenditures have tures. During the COVID-19 pandemic expenditures remained flat in nominal terms (Figure 5). related to the poverty-targeted programs CA and FSA actually fell.22 In 2018, one-in-five Serbians were covered by at least one social assistance benefit. Social assis- Spending on the two main poverty-targeted pro- tance benefits reached about 18 percent of the grams, CA and FSA, has given way to spending population,23 with the CA, FSA and the PA repre- on pro-natal benefits that are not means tested. senting the highest coverage (over 8 percent for the Significant reforms in child and social protection in CA and just under 4 percent for the FSA and PA). 2018 refocused spending on categorical pro-natal Whereas the number of CA household beneficiar- benefits (i) by increasing the eligibility and benefit of ies has fallen from nearly 210,000 in 2014 to a little wage compensation during maternity, childcare and under 137,000 in 2020, the number of PA benefi- extended childcare leave; and (ii) by making the PA ciaries increased from 62,000 in 2019 to 72,000 universal and substantially increasing its benefits, by September 2020. The number of beneficiaries especially for the third and fourth child (by a factor receiving support for childcare leave (wage com- of 5.2 and 5.7 respectively) (Figure 5). In 2019, the pensation) reached 42,200 in 2017, an increase of largest social assistance program was wage com- 1,500 since 2014. pensation during maternity, childcare and extended childcare leave. Such programs form an important Social assistance coverage of the poor is low and is part of social protection systems in other European significantly lower than comparator EU countries. countries; however, they are designed as social Social assistance covers less than half of the poorest Back to table of contents 9 22 UNICEF and World Bank (2021b). 23 Population living in households receiving at least one social assistance benefit. SERBIA Social Protection Situational Analysis Figure 5: Impact of child and social protection reforms on social assistance expenditures, 2013-2020 (Monthly expenditure by social assistance program, nominal terms) 3.5 3 2.5 2 Billion 1.5 1 0.5 0 Jan-13 Apr-13 Jul-13 Oct-13 Jan-14 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15 Jul-15 Oct-15 Jan-16 Apr-16 Jul-16 Oct-16 Jan-17 Apr-17 Jul-17 Oct-17 Jan-18 Apr-18 Jul-18 Oct-18 Jan-19 Apr-19 Jul-19 Oct-19 Jan-20 Apr-20 Jul-20 Oct-20 January Caregiver Allowance Child Allowance Pre-school subsidies Birth Grant Wage compensation FSA Source: Estimates based on administrative data provided by the government. Note: Expenditure data on wage compensation during maternity and childcare leave and other benefits related to childcare were only available from January 2019 onward. quintile and about 63 percent of the two poorest In theory, families and individuals may qualify for quintiles. While coverage of the poorest quintile multiple programs; in practice, there is very limited nearly doubled between 2014 and 2019, it repre- overlap and 77 percent of beneficiaries receive sents one half the coverage of new EU members support from only one program.27 such as Slovakia, Hungary, and Romania, where coverage approaches or exceeds 90 percent.24 Ser- Overly strict eligibility criteria, limited budget allo- bia’s categorical programs drive this gap either by cations and lack of transparent systems for iden- excluding the poor (for example, in the maternity tifying, selecting, and monitoring beneficiaries leave program) or by requiring but rarely carrying reduces social assistance coverage. CSW casework- out special outreach efforts to reach poorer house- ers have high discretion in determining whether holds. Categorical programs cover more than twice households’ actual or potential income makes them as many non-poor as poor children under age 5 (34 eligible for means-tested benefits.28 Together with versus 15 percent respectively). high caseloads, this results in arbitrariness in deci- sions and insufficient time dedicated to assisting Means-tested social assistance programs do not applicants or explaining reasons for rejection to sufficiently reach the poor. Only 3.5 percent of the concerned households. There is no monitoring and population receive the FSA, which is low when com- evaluation system with data collection and admin- pared to the share of the population living in abso- istrative oversight procedures to increase transpar- lute poverty (7.0 percent).25 Over half of the children ency and accountability. that belong to poor households are not covered. This is despite (i) an explicit government policy of The FSA program contains disincentives for labor supporting families with children; (ii) more than half market activation. Some of these disincentives are (56 percent) of social assistance spending allocated typical for programs guaranteeing minimum income to families with children;26 and (iii) a higher poverty as households just below the eligibility threshold rate among children than other age groups. More- have no incentive to earn additional income that will over, the average FSA and CA benefits amount be deduced from the support. By law, able-bodied to respectively 20 and 10 percent of the absolute FSA beneficiaries are required to register with the poverty line in Serbia (single adult) for the poorest National Employment Service (NES) and to partici- quintile, which is too low to meaningfully address pate in activation measures. However, the law is not the poverty. Low transfers from each program are clear on whether able-bodied FSA beneficiaries can offset by accumulating different types of benefits. participate in activation measures with remunera- Back to table of contents 10 24 Data from Social Protection Expenditure and Evaluation Database (SPEED) (2020). 25 Estimates based on data from MoLEVSA. 26 Refers to spending on the child allowance, parental allowance, wage compensation, foster care and preschool subsidy for 2020. 27 However, FSA beneficiaries are also entitled to other benefits such as, discount for utility bills, energy protected customer status, free meals in soup kitchens, free meals at school, reimbursement of the cost of preschool education, scholarships, one-off financial aid, and free transportation, among other, depending on the local government. 28 As per legal regulation to determine lost income, CSW is obligated to determine, for each household member, whether they are able to work and if not seek evidence (for example, medical reports), and identify what jobs, how many days a month of work and the expected earnings from that job. SERBIA Social Protection Situational Analysis tion (for example, wage subsidy schemes or public works). Moreover, the provision of services has been limited, due to limited institutional capacity within CSWs and NES. The fragmented institutional framework for iden- tifying beneficiaries and delivering social assis- tance undermines the efficiency of means-tested programs. The fragmented delivery mechanisms thwart coordination between CSWs and LSGs units, contribute to fragmented information systems and likely reduce access to benefits by more vulnerable groups. The government is currently developing an inte- grated social protection information system (SPIS), including a social card registry (SCR) to help strengthen efficiency and access. Notable recent progress includes the option to submit applica- tions electronically; for instance, a request for parental allowance can be submitted in the hos- pital via “E-baby.” The SCR, which is in place as of spring 2022, will enable better determination of the socio-economic status of applicants and their rights to benefits; increase the efficiency and trans- parency of the social assistance system; and enable the government to improve its shock response. The SPIS is not yet fully functional, and this may have limited the government’s response to COVID-19 (see below). Back to table of contents 11 SERBIA Social Protection Situational Analysis Social services Serbia’s social services system is being expanded with a broad system of programs and benefits and an increasingly diversified set of providers, including the central government, local self-governments, non-governmental organizations and the private sector. Expenditures on social services are compara- tively high by Western Balkan standards, but the provision of services is very uneven across LSGs, with many unable to serve growing needs in the population. High caseloads for the Centers for Social Work, fragmented delivery systems, and lack of standardized referral procedures limit coordination and effective services provision. Serbia has established the economic, social and many actors. The 2011 Social Welfare Law divides legal conditions for developing a system of high- social services into five main categories (Table 3). quality, sustainable, accessible and affordable These services are delivered across levels of govern- social services. With the 2011 Social Welfare Law, ment and through public and private providers, Serbia has been moving the system from an ad such as, civil society organizations, licensed profes- hoc development of services without a long-term sionals, private firms, and faith-based organiza- financing framework toward modern social service tions. Centers for Social Work (CSWs) are an integral provision centering on the family as the main unit. part of the system, responsible for: assessing indi- The legal framework also aims to foster a diversifi- vidual and family needs, candidates for guardians, cation of service providers and has, inter alia, intro- foster parents and adoptive parents; planning the duced a quality system that defines basic standards mix of services in their community; outsourcing and of social services and how these should be applied procuring third-party services; overseeing service for licensing, standards, and procurement of quality; and managing service providers (such as, services. therapy), even providers funded by the local govern- ment. Local self-governments (LSG) also manage, Service management and implementation is spread finance and provide services. between the central and local level and across Table 3: Social services by category and delivery agency Social services Delivery Assessment and planning: Assessment of a situation, needs, strengths CSW, Center for Family Accommo- and risks of users and significant others; assessment of guardians, foster dation and Adoption and Institu- parents and adoptive parents; development of individual or family plans for tion for Education of Children and service provision and legal protection measures. Youth Community services: Beneficiary support to remain in the immediate or LSG extended-family environment through the provision of services for day care, in-home assistance, temporary housing, among other. Independent living support: Beneficiary support for accommodation, LSG personal assistance, training for independent living and other types of support. Counselling and therapy, and social and educational services Mixed: LSGs and CSWs Accommodation services: Accommodation in a kinship, foster or other Mixed: LSGs and CSWs family for the adults and elderly; accommodation in the home for the elderly; shelter accommodation, and other types of accommodation Source: World Bank (2017) and 2011 Social Welfare Law. Back to table of contents 12 SERBIA Social Protection Situational Analysis Public providers dominate overall service provision youth with developmental and other disabilities whereas private providers face difficulties obtaining as well as elderly home care was low, especially in licenses due to lack of technical staff and unclear the European context,35 and poorer localities were procurement procedures at the local level. Between more likely not to provide (even this) social service. 2015 and 2018, the share of beneficiaries served by Services for independent living and shelters as the public sector decreased from 74 to 58 percent29 well as counselling, therapy and socio-educational and the share of non-governmental organizations services were available only in big cities. In general, (NGOs) increased from 26 to 35 percent. For-profit the LSGs with low service provision also allocated service providers, with a beneficiary coverage of 7 a relatively small share of budget to these services. percent, joined the market for the first time in 2018. Among public providers, CSWs accounted for 97 Public care for the elderly remains very limited percent of total service cases in 2015, although this despite an aging population. Most of the elderly partly reflects that CSWs oversee many low-in- who need assistance rely on personal networks. tensity services that have many users.30 Licensed Although this is the most common social service non-governmental providers mainly operate home provided, 2016 data show that only 1.5 percent of care and day care services for adults and the elderly, the elderly are covered by public institutional care, home care and personal assistance for children day care or home care services and that waiting lists and adults living with disabilities, and socio-edu- are long in larger cities. Private institutions are more cational31 services for children and young people. common but are unaffordable for the majority of NGOs account for about 35 percent of beneficiaries the population. Only 7 percent of the elderly receive for different home care or day care services and are the caregiver allowance, and the majority do not use the dominant providers of personal child attend- it to purchase care services.36 The shortage of resi- ants and personal assistance. dential and home services, coupled with free health care for the elderly, shifts the burden back into the Expenditures on social services are dominated health system and contributes to high and expen- by accommodation services. Public expenditures sive in-hospital care: it is common for the elderly to on social services represented 0.27 percent of be hospitalized for an extended period.37 GDP in 2018, of which a major share represented expenditures at the central government level (0.20 The central government is deinstitutionalizing percent of GDP).32 Two thirds of spending under the accommodation services as a key reform objec- national mandate (0.14 percent of GDP) was taken tive, but progress is uneven and quality problems up by residential and foster care services, with 0.06 persist. Caring for children in residential institu- percent for CSW staff costs (including assessment, tions as opposed to foster (family) care may have planning and partial operating costs). LSG financing negative impacts on children’s wellbeing and accounted for 0.07 percent for children with disa- development. In December 2020, the number of bility and personal attendants of children.33 children being accommodated was 5,823, of which 89 percent were placed in foster families and only There is a rising but unmet demand for social 11 percent in institutions. The number of children in services, and service availability varies significantly institutions decreased by 43 percent between 2011 across local government entities. The diversity and 2020, while the number of children in foster and number of users of social services is increasing families increased by 8 percent.38 However, children but coverage remains limited. According to the with more difficult conditions remain at much higher latest available data (2015), about 9 percent of the risk of institutionalization for life, and the quality of population was covered by social services in some foster care is still problematic with high turnover form, which was low compared to other European among families, inadequate care and low capacity countries.34 A review of social services delivered of many foster care families. The number of adults through LSGs also found significant variance and in institutions has increasing significantly between gaps in service provisions, and the new Law did not 2011 and 2020, respectively from 13,384 to 18,744. change the situation significantly between 2012 and 2018. The most common services provided (to Case management approaches and referral proce- more than 80 percent of beneficiaries) were elderly dures are not fully developed, limiting coordination home care (in 123 LSGs), day care for children with and service provision. In 2008, case manage- special needs (in 64 LSGs) and personal attendants ment was introduced in the CSWs. This systemic of children (76 LSGs). The coverage of children and approach in assessing activities, arranging access, Back to table of contents 13 29 Government of the Republic of Serbia (2018). 30 World Bank (2017). 31 Intensive support services for family in crisis, counselling and support of parents, foster parents, adoptive parents and similar. 32 Government of the Republic of Serbia (2018). 33 Ibid. 34 Based on case management and assumes each case is a new person; as this is often not the case, the share of population covered is likely overestimated by this approach. 35 For example, for a set of EU countries, the average coverage of elderly by LTC home-based care was 7.3 percent in 2016 (OECD 2020). 36 European Commission (2018). 37 Hirose and Czepulis-Rutkowska (2016). 38 Government of the Republic of Serbia (2020a, 2021). SERBIA Social Protection Situational Analysis planning, coordination and evaluation is still not fully implemented, however. Institutional resistance partly stems from high caseloads and extensive reporting requirements. Clear criteria and proce- dures for referrals of actual and potential clients for services (or social assistance) are lacking between CSW and LSG social welfare departments or other involved entities. Staff shortages and skills constraints limit CSW capacity to deliver effective and high-quality services to the target population. The CSW work- force fell by 18 percent between 2014 and 2018, due to a ban on employment in the public sector in 2013.39 The caseload has consequently increased, reducing the capacity for tailored, client-oriented services. There is no comprehensive skill-upgrading strategy, and trainings are provided on a limited and ad hoc basis. Gaps in staff capacity are particularly serious in relation to prevention of and protection against violence and work with victims of violence and with perpetrators. Local governments cite lack of funds, staff and competences as major reasons why some social services are not offered.40 Ear-marked financing from central to local level is intended to boost LSG service provision, but weaknesses related to targeting and substitu- tion effects may reduce its effectiveness. Trans- fers are earmarked for services provided by local governments in less-developed areas (80 percent of resources), those that have institutions for resi- dential care undergoing transformation or that are developing innovative and important social services. However, some eligible LSGs represent relatively well-off municipalities - only the most developed cities and municipalities are ineligible. A compar- ison of LSG services between 2015 and 2018 (the transfer policy came into effect in 2016), suggests that LSGs in general did not replace existing services with national financing but used transfers to develop services. At the same time, the financing of LSG social services as a percentage of GDP did not change between 2015 and 2018; and in almost half of LSGs, local budget allocations excluding transfers either decreased or remained at zero, suggesting that at least some LSGs simply replaced, rather than expanded, financing on social protection.41 Back to table of contents 14 39 Government of the Republic of Serbia (2021). 40 Government of the Republic of Serbia (2020b). 41 Ibid. SERBIA Social Protection Situational Analysis Employment and labor market policies Serbia spends comparatively little on unemployment benefits and even less on active labor market programs (ALMPs). Coverage of both passive (unemployment benefits) and ALMPs is low by international standards. Among ALMPs, the offer of training programs is limited and largely focused on on-the-job training for recent university graduates rather than inclusion of more vulnerable populations or reskilling of adults needed to address skill gaps and modernize the workforce. The efficiency and effectiveness of labor market policies is held back by lack of human resources, technology and information tools that could increase efficiency and relevance of day-to-day work as well as long-term strategy development. Labor market programs are managed by Serbia’s tional elementary education of adults), life-skills National Employment Service (NES). The NES, training, support for self-employment programs, a public agency under the supervision of the wage subsidies for hard-to-employ jobseekers, Ministry of Labor, Employment, Veteran and Social incentives for FSA beneficiaries and public works. Affairs (MoLEVSA), has three key responsibilities: (i) administering unemployment benefits (passive Expenditures on passive and active labor market labor market programs); (ii) supporting the unem- programs are low in comparison to more developed ployed through job-counselling, intermediation, economies in Europe and the OECD. The budget and provision of ALMPs; and (iii) supporting Serbia’s envelope for NES reached RSD 22,600 million (USD lifelong learning system with skills needs forecasts. 215 million) in 2019. In response to COVID-19, the NES caseworkers set up profiles, determine eligi- budget was increased by almost 20 percent in bility for different types of support and provide nominal terms in 2020, to RSD 26,705 million (USD counselling and job-search assistance. NES offers 256 million), reflecting higher spending on both different ALMPs to the unemployed, based on unemployment benefits and ALMPs. Yet, Serbia needs assessments carried out by the caseworkers. generally spends very little on ALMPs, about 0.08- Following the World Bank classification, ALMPs 0.09 percent of GDP between 2017 and 2019. EU provided by NES include employment mediation unemployment rates average below those of Serbia, services, job-search assistance (such as, job-search but the average share of GDP allocated to expendi- training, counselling, job-search monitoring, job tures on labor market programs is significantly higher clubs, and job fairs), skills development (mostly for both passive and active programs, representing on-the-job training), basic skills training (func- 0.63 and 0.38 percent of GDP, respectively (Figure 6). Figure 6: Share of passive and active labor market programs, 2019 1.4% Total LMP (incl. administration, placement and related services) 1.2% Active LMP (including Vocational Training) 1.0% Passive LMP (unemployment benefits) 0.8% 0.6% 0.4% 0.2% 0.0% EU-27 7STEE Montenegro (2019) BiH North Macedonia Serbia Albania Sources: EU-27/7STEE: Eurostat; North Macedonia: Administrative Data; Albania: Administrative Data NAES. Montenegro: MONSTAT and ZZZCG. Note: 7STEE refers to the seven small transition economies of Europe, namely Bulgaria, Croatia, Estonia, Latvia, Lithuania, Slovak Republic and Slovenia. Back to table of contents 15 SERBIA Social Protection Situational Analysis In the labor force survey (LFS), there is limited between 2019 and 2020 reflected a youth-oriented overlap of registered unemployed and actual program entitled My First Salary. This program jobseekers, suggesting that many registered in the aims to provide the first labor market experience NES database are not actively seeking employ- to about 8,000 youth with at least secondary ment. In 2020, the number of total NES-regis- education; the first cohort started in December tered unemployed averaged 509,000. This number 2020. Another workplace-based trainee program exceeded, by more than 200,000, the number of (Professional Traineeship Program) also places unemployed using LFS data and the ILO defini- educated youth in jobs. In contrast, a private- tion of active jobseeker.42 At the same time, only 64 sector wage subsidy (incentive) program is more percent of those reporting to be unemployed in the representative of the stock of registered unem- LFS were registered with the NES.43 The gap could ployed, covering largely middle-age unemployed consist of those informally employed who wish to people with a secondary school education. Public retain access to NES services.44 It could also indicate works programs are more likely to engage men and a lack of updated profiles. Between 2020 and 2021, participants with lower educational backgrounds. the number of unemployed fell by some 20 percent, reflecting the lack of impact of COVID-19 on official Despite significant improvement, the caseload of unemployment rather than restricted access to NES caseworkers is still too high for effective case NES services during the pandemic (see below). management. In recent years, NES has rational- ized its caseloads by diminished the regional vari- Of the total registered unemployed, a small minority ation across branch offices.46 In 2020, the average receive an unemployment benefit. The caseload was 816 jobseekers per caseworker, many unemployment benefit is available to those who times higher than the ILO recommendation of are covered by unemployment insurance and around 100. High caseloads reduce the quality of fulfill additional standard eligibility NES “gateway” service, such as profiling and needs requirements.45 The unemployment benefit is analysis, which are the basis for other services. claimed by only 6 percent of the registered NES does not use technical tools (such as, statis- unemployed, reflecting a lack of formal and tical profiling) to save case management time and longer-term employment opportunities, the frequently update and adapt systems. A job-portal long-term unemployed who no longer meet eligi- function providing job-matching services is not bility criteria and a disproportionate share of young widely used. Programs are not systematically eval- first-time jobseekers among the unemployed. uated, and, with a few exceptions, no causal impact evaluations are made to determine value-added Active labor market programs have few partici- and cost-effectiveness. Standardized performance pants. The job-search assistance uptake is universal, metrics such as net placement rates are not yet used. as meeting with a caseworker to register a profile is a requirement for registering as unemployed The system to develop and deliver training and regular meetings are required to remain in the programs aimed at providing new or comple- register. The participation in other and non-com- mentary technical skills is underdeveloped and pulsory forms of activation programs is very limited, inflexible. In NES, only one skills-training program however. Skills development training participants is focused on new technical (vocational) skills. In numbered 10,000 in 2019 and 15,000 in 2020. 2021, the NES National Employment Action Plan47 Private sector incentive schemes (such as, entre- changed to a three-year period (2021-2023) to preneurship programs and wage subsidies for new implement the National Employment Strategy jobs) and public works reached 2 and 1 percent (2021-2026). The strategy defines employment respectively of NES unemployed clients in 2020. policy goals for ALMPs and annual procure- ment of services for external training providers to Skills-training programs are focused on providing deliver ALMP measures (such as, short programs workplace experience for youth, often college for adults). These technical trainings are procured graduates. Youth, university graduates and those by NES from external contractors. However, the who have been looking for a job for less than one number of providers is limited, as accreditation (of year are overrepresented among the skills training providers and courses) by the Ministry of Education, beneficiaries, compared to their share in the total Science and Technology Development (MoESTD) number of registered unemployed. Other programs is required but difficult to achieve. Complex and are balanced relative to the distribution of unem- costly procedures discourage providers from ployed. The increase in skills development programs Back to table of contents 16 42 However, the definition of active jobseeker in the Labor Force Survey might be stricter than what NES uses to register unemployed individuals. According to the law, NES must maintain in its records those individuals who have not established employment or otherwise exercised the right to work, while the survey identifies as active jobseekers those who actively sought work during the four weeks have preceded the observed week and are able to start work within two weeks after the end of the observed week. 43 Information drawn from 2020 Labor Force Survey data. Refers to population 15 years old and above (SORS 2021). 44 Before 2007, access to means-tested free health insurance was conditional on registration with NES. As of 2017, this is only the case with employment termination. 45 Employed (including temporary) or self-employed, (i) insured over the previous 12 months without breaks or for 12 months within the previous 18 months with breaks; (ii) employment was terminated without employee’s will or fault; and (iii) employee was declared redundant and did not receive severance payment from the Budget of the Republic of Serbia in excess of the amount of severance payment determined by the Labor Law. 46 In 2016, the ratio of unemployed to counselors ranged from 600 to over 2,000 across regions (World Bank 2018). 47 The National Employment Action Plan is adopted by the government at the proposal of the MoLEVSA, and NES is one of the implementors of activities. SERBIA Social Protection Situational Analysis seeking accreditation or develop innovative courses personal income tax, reduce incentives for formal and programs that meet changing skills develop- employment at the lower wage end of the spectrum ment needs. Moreover, the procurement process is (that is, for groups where activation is most acutely lengthy, complex and repeated each year even for needed).48 Serbia also lacks a clear strategy for active trainings that are organized in similar ways, and the ageing—including end-of-career psychology and approach to evaluating training quality in proposals management, health care directed at older workers has not been standardized. A new public procure- or opportunities for combining pensions and ment law addresses these weaknesses, including work—that would prolong an active working life.49 raising limits for public procurement and simplifying steps. NES is also in the process of reviewing best practice in other European countries for effectively defining and evaluating training quality criteria. Lifelong learning systems to reskill working and unemployed adults are underdeveloped and uneven. Despite low productivity, an aging popu- lation and the competitive pressures from global skills dynamics, opportunities for lifelong learning and prolonging productive working life are only nascent. Slow training course and provider accred- itation processes, as well as lack of an advanced labor market information system to guide the development of training, limit progress. NES is mandated to set guidance for national skills devel- opment in its annual plan. However, lack of an integrated and automated data management system (see below) limits needs assessments, training opportunities and sectoral evaluations of annual data. The policy coherence between key lifelong learning systems (that is, education and national qualifications frameworks) is also weak, undermining coordination across different actors. Serbia’s labor market information system (LMIS) needs to become a more advanced and integrated system to support NES functions and improve labor market functions. The NES information system forms the basis of the LMIS and is a modern, operational system that provides high-quality data. The NES integrated systems provide some basic LMIS services to case managers and clients, such as job-search assistance and job matching. However, the information does not inform decision making, program monitoring or program evaluation. Nor is this information used to provide data analytics for users external or internal to NES, or to direct online career/skills development guidance services for individuals or enterprises. NES has not yet adopted available technology tools that permit high-frequency analysis based on real-time data. Work incentives need to be addressed. Currently, high social security contributions, together with the Back to table of contents 17 48 World Bank (2019). 49 Jakovljevic, Jovanovic, Milovanovic and Radevic (2020). SERBIA Social Protection Situational Analysis Pension system A rapidly aging and shrinking population is putting significant pressures on the Serbian public pay-as- you-go pension system. Reforms implemented between 2014 and 2018 helped stabilize the fiscal costs of pensions, but at the price of lower adequacy. Strong labor market developments in recent years (up until Covid-19) increased the number of contributors in the system. However, these trends are not likely to be sustained over the long term. Almost one-in-ten Serbians aged above 65+ is not covered by pension schemes, and beneficiaries can only expect to replace half of their income as retirees. The poor are also less likely than non-poor to receive pensions, as is characteristic for contributory systems. Going forward, trade-offs between coverage, adequacy and sustainability will become even sharper. As a result, Serbia needs to review the parameters of its pension system, foster higher savings among those currently employed and consider non-contributory options for the elderly poor. The Serbian pension system is dominated by the Contributory revenues account for the major share public pay-as-you-go (PAYG) system. The admin- of revenues, reaching 8.7 percent of GDP in 2020. istration of the public pension system is integrated Due to increases in total employment and average under a single public entity, the Pensions and Disa- wages, contributory revenues steadily increased bility Insurance Fund of Serbia (PIO). The Law on between 2012 and 2019, before declining in 2020. At Pension and Disability Insurance of 2003 intro- the same time, the cap on pension outlays reduced duced the point structure in the system,50 reduced budgetary transfers from 7.1 percent of GDP in 2012 the number of minimum pensions to one, increased to 3.1 percent of GDP in 2019. Contributory revenue retirement ages, introduced rule-based indexation reached 9.1 percent of GDP in 2019 before falling and changed the eligibility for the disability pension. back to 8.7 percent of GDP in 2020. As a conse- Pension legislation has subsequently been subject quence of the negative shock to contributions due to numerous changes, including for retirement age, to the pandemic, budgetary transfers increased contribution rates, indexation formula and value of substantially to 3.9 percent of GDP (30 percent of minimum pensions.51 In addition, the government total revenue) in 2020, from 3.1 percent of GDP has introduced ad-hoc pension increments in certain or 25 percent of total revenue in 2019. However, years. The pension system, nonetheless, maintains the share of budget transfers was reduced to 14.8 the defined benefit design, PAYG financing and percent of total transfers in 2021.55 By contrast, the administrative structure that existed under the farmers pension is primarily (90 percent) paid with Yugoslav regime prior to 1990. The pension system state budget (Figure 7). includes a pensions scheme for farmers, which is semi-contributory in nature owing to its structure Diverging trends among beneficiaries and contrib- based on flat contributions and flat benefits.52 utors have resulted in a low support ratio that threatens the performance of the system, while the Demographic and employment trends continue to more vulnerable have less coverage. The rapid aging put pressure on the sustainability of the pension of the population, fueled by low birth rates and high system financing. Spending on pensions accounted outmigration, is not compatible with the current for 10.3 percent of GDP in 2020, a significant PAYG set-up. At 1.3, the ratio between contribu- reduction compared to 12.3 percent of GDP in 2014. tors and pensioners (the support ratio) in Serbia is The sustainability of the pension system, despite a now one of the lowest in Europe.56 Recent reforms temporary reform to avert a crisis in 2014-2018,53 that increased retirement ages have reduced the remains an issue. Population dynamics bring fewer number of beneficiaries, concentrating them in the working adults and more elderly, and the share of 65+ age category, thus increasing coverage among informality in employment is increasing. Higher the elderly. However, this trend is unlikely to be informality lowers the share of contributors among sustained in the future in the absence of additional the employed, which stood at 75.6 percent in 2019.54 reforms or structural changes in the labor market. Back to table of contents 18 50 Workers earn pension points based on their individual earnings with respect to national average earnings for each year of contributions. At retirement, the sum of pension points is multiplied by a pension-point value to convert them into a regular pension payment. 51 Pension system parameters are in Annex 2. 52 A special cash benefit has also been introduced for parents reaching retirement age who are not entitled to a pension but who have personally cared for a child with severe disability for a minimum of 15 years; this cash benefit emanates from the social assistance system, however, and is claimed by only 600. 53 Fiscal consolidation (implementation of the Law on Temporary Regulation of the Manner of Pension Payment, which reduced pensions above the determined amount) was conducted in the period from November 2014 to September 2018. 54 The share increased to 78 percent in 2020, but this may reflect the impact of temporary employment policy actions due to Covid-19. 55 Estimates for 2021 indicate that pension expenditures might have decreased, and contributory revenue might have increased once again. Definite figures are not available yet. 56 Although this figure represents a small increase compared to the 1.1 level observed in 2014. SERBIA Social Protection Situational Analysis Figure 7: Revenues and expenditures, 2012-2020 16% 14% 12% 7.1% 6.3% 6.0% 10% 4.9% 4.5% 3.9% 3.4% 3.1% 3.9% % of GDP 8% 6% 4% 8.3% 8.4% 8.3% 8.5% 8.7% 9.1% 8.7% 7.5% 7.8% 2% 0% 2012 2013 2014 2015 2016 2017 2018 2019 2020 Social contributions Transfers from the budget Pensions Expenditures Total Expenditure Source: Pension Insurance Fund of the Republic of Serbia (2012-2020). Data from 2015 indicate that 12 percent of pension observed in 2019, employment rates, the share of benefits go to the poorest 20 percent of the popu- formal employment and wages would all have to lation, while 29 percent of benefits are paid to the increase by a combined 15 percentage points at richest quintile. The farmer’s pension, whose bene- least. This would be extremely challenging. Deliv- ficiaries largely comprise lower income households, ering more and better pension benefits will require covers just over 160,000 people.57 There is also a raising pension expenditures. Given the PAYG set-up, disability pension, covering about 266,000 people reforms to the valorization of wages would increase in 2020.58 replacement rates with high fiscal costs. A balance should be struck to provide pensions in payments Serbia’s reforms to address the sustainability of with a less generous adjustment mechanisms, while pension financing have come at the price of lower still protecting against inflation. Despite an overhaul adequacy. A fiscal consolidation program in 2014 in 2014, favorable conditions for early retirement reduced the value of pensions above a certain level, also still apply for several professions. The decline which affected almost 40 percent of pensioners. in the farmer’s pensions system indicates that this Between 2015 and 2018, average pensions saw very system is losing relevance, while there is no social small adjustments and the average replacement pension for elderly poor. rate (calculated as the average pension divided by the average pre-retirement income observed in the Voluntary schemes, meanwhile, are underdevel- system) for old-age pensions declined to 50 percent oped. The coverage of voluntary pension plans has in 2020, down from 61 percent in 2012.59 When stagnated around 10 percent of total employment measured as the ratio between average pensions and with assets equal to 0.85 percent of GDP. The and average gross wages, benefit adequacy is higher share of active members (defined as those who in Serbia than in other Western Balkan countries but paid contributions in the last year) declined from lower than in most European countries and below 4 percent in 2010 to 3 percent of employment in the EU average. Adequacy is also reduced through 2020. High labor mobility and high insecurity in the current approach for valorization of wages labor markets reducing job tenure undermines the and indexation of pensions. Under current rules, system. Incentives to participate are also limited: past wages are actualized at less than the overall voluntary retirement contributions are exempt from wage growth in the economy, affecting the value taxes up to a level equivalent to about 8 percent of benefits. Pensions do make an important contri- of average wages in Serbia, which is also quite low. bution to the income of the poorest quintile that International evidence suggests that tax incentives receive pensions. For this group, pensions account are not an effective mechanism to increase partic- for 64 percent of total income. ipation in retirement savings in the middle class as incentives are highest for high-income individ- Ultimately, reductions in coverage and adequacy uals and workers closer to retirement. Develop- are threatening old-age income security in Serbia. ment of occupational (employer-based) pension In the next 20 years, the working-age population plans with auto-enrolment, financial education of is expected to decrease by more than 15 percent, plan members, among other, have proven effective while the elderly population is expected to increase elsewhere but are not yet implemented in Serbia. by more than 8 percent. To maintain revenue levels Back to table of contents 19 57 Pension and Disability Insurance Fund of Serbia (2021). 58 Ibid. 59 In the period of application of the Law on Temporary Regulation of the Manner of Payment of Pensions, in accordance with fiscal possibilities, pensions were adjusted, as follows: (i) pensions for December 2015 increased by 1.25 percent; (ii) for December 2016 increased by 1.5 percent; and (iii) for December 2017 increased by 5 percent. SERBIA Social Protection Situational Analysis Lessons learned from Covid-19 The government responded quickly to the COVID-19 pandemic with an economic stimulus package that successfully protected output and jobs. Policy measures also centered on sustaining access to public services, including those of social protection such as social services and NES services, while containing the risk of contagion. Notwithstanding these accomplishments, the social protection system failed to respond effectively to the crisis. Expenditures on social protection did not increase substantially in response to the shock. The social assistance system, which has a particularly important role in protecting the poor and vulnerable, was not leveraged by either increasing coverage or adequacy of benefits to poor and vulner- able. The two one-off cash transfers provided were universal (all adults) and semi-universal (to pensioners). They consequently proved expensive but had limited impacts on poor households, poor children and the elderly poor. NES managed to reorganize services online, but apart from a trainee program for youth, active labor market programs were not restructured to address new challenges. The government reacted rapidly to minimize the beneficiaries (for example, those benefitting from economic impact of the COVID-19 pandemic and social services). Social assistance recipients’ entitle- successfully protected output and jobs. Resources ments were automatically renewed, and electronic were reallocated to mitigate health impacts, protect applications were encouraged. Some LSGs provided the elderly and provide remote learning, and two assistance packages and NGOs contributed to universal cash transfers were provided. Economic supporting the Roma. growth fell in 2020 but only by one percent,60 and employment and unemployment rates have The social assistance system does not have the remained stable so far. Poverty rates were only legal basis or capacity to expand horizontally or marginally affected i n 2020;61 y et, a s economic vertically to respond to shocks. The Serbian social and employment stimulus measures focused on assistance system is set up for idiosyncratic shocks preserving formal private sector activity (especially and lacks a policy for systemic emergencies. The small and medium enterprises) and employment, SPIS and the SCR are still being developed; they those not formally employed including informal are key components to rapid systemic scale-up. workers and farmers, children, and the elderly were Whereas there are legal provisions for expanding marginalized. Declining access to public services the FSA in times of need, program eligibility is based including social services is likely to have dispro- on identifying people in chronic, long-term poverty portionately affected v ulnerable populations, t he (that is, highly restrictive, see above) and it remains a longer-term effects of which may not yet be visible. paper-based application process (that is, unsuitable for a shock). However, for those social assistance The Covid-19 response elicited a rapid response beneficiaries whose right to social benefits expired from institutions in charge of social protection, on 15 March 2020 and later, the government has but actions centered on maintaining services extended their entitlements for a maximum of three despite the health threat rather than expanding months (that is, during the state of emergency).62 social protection. Within social protection, social assistance addresses systemic shocks that The coverage of social assistance programs did threaten the poor and vulnerable. The specific nature not increase during the first year of the epidemic; of the COVID-19 shock and the risk of contagion instead, two costly untargeted transfers were raised challenges of delivering public services and provided. The government provided one-off cash benefits. Regulations adopted by MoLEVSA aimed assistance to all adults (EUR 100, approx. to mitigate the negative impacts and ensure the RSD 12,000 or USD 115) and a one-off payment continuity and safety of the system. This included to all pensioners (RSD 4,000 or USD 38). These measures to protect social protection workers and transfers did not target the poor and vulnerable, especially Back to table of contents 20 60 World Bank (2021b). 61 World Bank (2021b). 62 The government extended benefits for social assistance beneficiaries (FSA, Caregiver Allowance, augmented [This is not mentioned in the text. What is it?] Caregiver Allowance, Child Allowance, paid leave for special childcare and other entitlements related to special childcare). Government Conclusion 05, No 53-2787/2020 of 24 March 2020. Available at: https://www.srbija.gov.rs/extfile/sr/454269/zakljucak_socijalna_zastita234_cyr2.zip SERBIA Social Protection Situational Analysis not children, the group most likely to live in poverty to COVID-19 challenges, most OECD countries in Serbia but also most likely to live in large house- adapted classroom based ALMP training to online holds with many children, reducing the per capita training modes and also developed additional online value of the transfer. Unlike FSA recipients, moreover, training programs. Moreover, several countries recipients of the child allowance did not automati- also increased the number of training places and cally receive the cash transfer but had to apply for provided short online job training in professions in it.63 The total direct cost for the extraordinary cash demand (including care workers).65 transfers amounted to EUR 673 million (USD 730 million), which is equivalent to about 73 percent of total social assistance expenditures, and over three times the level of expenditure of the means-tested programs. Social services were initially interrupted, affecting the vulnerable elderly and persons with disabilities. During the state of emergency home care services were interrupted in many places due to caregiver shortfalls and public transport restrictions. In many municipalities, caregivers no longer entered homes but merely deposited food and medicine outside. Day care and child attendant services were likewise discontinued during these months. Most LSGs met these challenges by organizing call centers and delivery services to collect and fulfil requests for food and medicine. Call centers also provided infor- mation and psychological support services. The impact on the pension system is not yet clear. Aside from the one-off transfers and as part of the enterprise support package, the government estab- lished a three-month deferment of social security contributions for all private companies, to be repaid as of 2021. Pension finances will be affected by this and the reduction in contributions that followed on the economic downturn, but it is not yet clear what the long-term impact will be. The NES managed to maintain continuity in basic service provision and to scale-up or reform the ALMPs to some extent. Registration and benefit claims as well as job-matching services were trans- ferred online.64 The registration of unemployed dropped during lockdowns but did not surge as restrictions were lifted, supporting the general conclusion that formal jobs were protected from the crisis. Some ALMPs were adapted to secure continuity. NES also launched a new program for graduates (My First Salary) and organized the First Virtual Employment Fair in 2021. In 2022, the NES is launching an online training platform to replace in-person trainings including on entrepreneur- ship and job search for new graduates. However, NES did not reform the training offer substantially. An overview by the OECD shows that in response Back to table of contents 21 63 Government of the Republic of Serbia (2020c). 64 NES still allows use of email to submit different types of requests; however, as of the second quarter of 2021, all services were again offered in-person. 65 OECD (2021). SERBIA Social Protection Situational Analysis Summary and reform priorities Serbia’s social protection system should be oped information systems reduces the opportuni- strengthened to increase resilience against shocks ties for client-oriented labor market services. The and to help protect its population against poverty. evaluation of ALMPs has been enhanced in the past The social protection system in Serbia is compre- period, but it needs further improvements. hensive and complex, spanning many objectives such as reducing vulnerability and poverty across Despite reforms, the sustainability of the pensions the lifecycle, supporting families with children and system remains precarious. Pensions account for promoting higher birthrates, and promoting access the largest share of public expenditures on social to more and more-productive employment. The protection. The coverage and adequacy of the analysis above shows, however, that reforms are pension system is falling, however, and aging trends needed to increase coverage and effectiveness, will continue to put pressure on the system. At the strengthen resilience and preserve financial sustain- same time, measures to foster complementary ability. retirement savings are ineffective, and there are not specific measures for the elderly poor. Social assistance is increasingly focused on pro-natal categorical programs, reducing resources The COVID-19 pandemic also exposed weaknesses available for protecting the poor and vulnerable in the social protection system. Serbia’s govern- through means-tested programs. By construction, ment reacted swiftly and decisively to contain the spending on pro-natal programs also risks escalat- COVID-19 crisis and protect formal employment. ing to the point where they absorb or exceed the However, neither information systems nor a legal entire envelope for social assistance expenditures. framework was in place to facilitate an upscaling Access to poverty-targeted programs is held back of the social assistance system, by either increasing by overly strict eligibility criteria. As a result, too coverage or adequacy of benefits to poor and vul- many poor, including children, are left uncovered. nerable. The one-off cash transfers provided did not target poor households, poor children, or the elderly. Serbia is developing the offer of social services, but progress is uneven across the LSGs reflecting gaps These findings point to a substantial reform agenda in human and financial resources. Recent informa- to increase sustainability, strengthen effective- tion is lacking on the state of social services in Serbia. ness in reaching the poor and develop information A review of LSGs shows that provision varies signifi- systems to support policy development. Among cantly and that critical services such as shelters, pre- others, the government could consider interna- vention and protection against violence and work tional experience on best practice approaches for with victims of violence are often lacking altogether. achieving pro-natal objectives, while redirecting Limited outreach results from lack of resources and social assistance spending toward mean-tested staff capacity resulting in high caseloads, and frag- approaches to ensure coverage of the poor and vul- mented delivery systems with poor coordination nerable. To increase financial sustainability, there is across social assistance and social services. As a also a need to continue to adapt parameters of result, access to social assistance and social services the pension system, encourage voluntary savings is conditioned on where a person lives and not (only) and develop policy incentives to prolong the on a person’s level of vulnerability. working life. To increase outreach and effectiveness, institutional coordination and Expenditures on employment programs, in particu- harmonization of service provision will need to lar ALMPs, is limited and coverage of the unemploy- improve significantly, and access-friendly ment benefit and ALMPs is low. High caseloads, approaches such as one-stop-shops for lack of time-saving technology and underdevel- social assistance, social services and employment programs should be implemented. Back to table of contents 22 SERBIA Social Protection Situational Analysis At the same time, the transition toward case man- results-based management, promote coordination, agement approaches needs to be sped-up across help clients and staff identify adequate forms of different institutions. More investment is needed in support, and facilitate access and increase trans- client-oriented services by reducing caseloads and parency. Finally, Serbia should review the experi- upskilling staff and by capacitating less developed ences from the COVID-19 response and establish LSGs in providing social protection services to its the necessary policy framework to leverage the population. Efficiency gains are needed across the social protection system in times of crisis. The board to free up resources for these reforms. Serbia proposed policy recommendations and knowledge should continue to build up the information systems gaps are outlined in more detail in Table 4. and digital tools that assist in system oversight and Table 4: Policy recommendations and knowledge gaps Short-term reform Medium- to long-term reform Knowledge gaps Social assistance • Reconsider the objectives of and • Establish a one-stop-shop for social • Promote more effective coordi- spending on categorical benefits and assistance and social services to facil- nation between the Center for make adjustments to improve these itate access and increase efficiency, Social Work (CSW) and the Local benefits. including for harmonized means Self-Government (LSG). • Shift spending away from categorical testing. programs and means test social assis- • Harmonize means testing for different tance programs or introduce means- targeted benefits and improve the testing for categorical programs (such capacity of means testing in LSGs. as the Family Allowance). • Continue to develop and implement • For means-tested programs, clarify the Social Protection Information and relax eligibility criteria and limit System (SPIS) and the Social Card discretion in the treatment of posses- Registry. sion of movable assets. • Integrate social assistance and • Improve outreach efforts for labor market services to promote poverty-targeted programs with a activation, to reduce dependency particular focus on rural areas and and to remove work disincentives in households with children. social assistance programs; promote in-work benefits. • Subject to budget availability, increase benefit amounts in means- tested programs. Social services • Accelerate the de-institutionalization • Continue to expand public and • Map social services: coverage, reform. private provision of social services. expenditure, adequacy, main • Create a one-stop-shop across • Revisit eligibility criteria for earmarked delivery institution, among other. delivery systems: one location, single transfers. • Map vulnerable populations who contact point, single entry point for • Establish an integrated information could be targeted for services. applications, case management, and system to monitor vulnerable popu- • Adhere to best practice of service payment functions. lations. delivery, including the balance of • Reduce caseloads and ensure • Review options for informal long- remedial versus preventive services adequate staffing and clearer division term care, including the relevance of and areas for improvement. of labor in CSW. Caregiver Allowance against vouchers • Enforce case management or paid care leave, and consider approaches and training for CSW training for informal (family) care staff. workers. • Develop approaches to coordi- nate and combine social services depending on community needs and service availability. • Improve transparency and awareness by increasing the scope and quality of information for prospective service users. Back to table of contents 23 SERBIA Social Protection Situational Analysis Short-term reform Medium- to long-term reform Knowledge gaps Pensions • Continue to revisit pay-as-you-go • Develop policies on active aging and • Analyze the adequacy of pensions (PAYG) system parameters and adjust delayed exit from working life. and determine what percentage of to improve financial viability and • Gradually phase out the Farmers consumption does pension income adequacy. Pension and introduce a social represent for the elderly? • Review generosity of early retirement pension to reduce the risk of poverty • Carry out a redistributive analysis categories. in old age. Relocate some fiscal costs of pensions and implicit subsidies. • Foster complementary savings mech- associated with the Farmers Pension. • Analyze the impacts of increasing anisms such as: the retirement age on the labor o occupational pension plans; force participation of older workers. o voluntary savings plans; and • Determine the characteristics of o voluntary low-cost interventions participants in voluntary pension appealing to self-employed. schemes by size of employer, indi- vidual earnings, etc. • Determine the poverty impacts of alternative designs of non-con- tributory benefits targeted to the elderly and persons with disabil- ities. Employment and active labor market programs (ALMPs)66 • Increase the efficiency of procurement • Increase government resources for • Review of ALMPs for coverage, of quality private ALMP providers. ALMPs. beneficiaries, objectives, outcomes • Establish rigorous monitoring and • Continue to develop labor market and impact evaluations, among evaluation (M&E) procedures and information system (LMIS). other. data systems based on best practice. • Integrate social assistance and labor • Review the incentive structure for • Systematically align ALMP offer with market services to promote activation work vis-a-vis the tax wedge. information on labor markets and and to reduce dependency; remove jobseeker characteristics and include work disincentives in social assistance long-term skill needs perspective. programs; and promote in-work • Reduce the NES caseload burden by: benefits. o updating the beneficiary registry; o reducing administrative tasks for caseworkers; and o implementing technical tools such as statistical profiling or information systems analytics. Disaster Preparedness • Develop lessons learned from the • Continue to build the SPIS and LMIS COVID-19 response. to develop sources of high-frequency • Develop strategy on remedial action data. for vulnerable groups hardest hit by • Continue to develop the SCR as the COVID-19. key policy tool for social protection • Launch the Social Card Registry shock response. (planned). • Establish disaster risk protocols to allow the FSA (or other key programs) to expand coverage in response to shocks. • Identify financing mechanisms, including the reserve fund and budget reallocation. • Map (internal/external) actors respon-sible for contributing to the disaster response. Back to table of contents 24 The employment strategy for the period from 2021 to 2026 addresses many of these recommendations and is available at: https://www.minrzs.gov.rs/sr/dokumenti/ostalo/ 66 sektor-za-rad-i-zaposljavanje, in Serbian and English. SERBIA Social Protection Situational Analysis List of References Cuberes, D., and M. Teignier. 2016. Aggregate Jakovljevic M., M. Jovanovic M., O. Milovanovic and Effects of Gender Gaps in the Labor Market: S. Radevic. 2020. “Serbia.” In: Ní Léime Á. et al. A Quantitative Estimate. Journal of Human (eds) Extended Working Life Policies. Springer, Capital 10(1):1–32. doi:10.1086/683847. Cham. https://doi.org/10.1007/978-3-030- 40985-2_31. European Commission. 2018. ESPN Thematic Report on Challenges in Long-Term Care: Organization of Economic Co-operation and Serbia. Development (OECD). 2021. Active Labour Market Policy Measures to Mitigate the rise in Eurostat. 2021. European Labour Force Surveys (Long-Term) Unemployment. A Summary of Database. Available at https://ec.europa.eu/ Country Responses to the OECD-EC Ques- eurostat/web/lfs/data/database. tionnaire. Paris: OECD Publishing. Government of the Republic of Serbi. 2021. Report Organization of Economic Co-operation and on the Work of Centers for Social Work. 2020. Development (OECD). 2020. Spending on Social Protection Institute. Belgrade. Available long-term care. Policy Brief. Paris: OECD at: http://www.zavodsz.gov.rs/media/2159/ Publishing. izvestaj-o-radu-csr-2020.pdf Organization of Economic Co-operation and Government of the Republic of Serbia. 2020a. Development (OECD). 2019. PISA 2018 Report on the Work of Centers for Social Work. Results, Volume I: What Students Know and 2019. Social Protection Institute. Belgrade. Can Do. Paris: OECD Publishing. Government of the Republic of Serbia. 2020b. Pensions and Disability Insurance Fund of Serbia. Mapping Social Care Services within the 2021. Yearly Statistical Bulletin 2020. Belgrade. Mandate of Local Governments in the Republic of Serbia. Belgrade. Robayo-Abril, M., and N. Millán. 2019. “Breaking the Cycle of Roma Exclusion in the Western Government of the Republic of Serbia. 2020c. Balkans.” Washington DC: World Bank Group. Social Safety Nets in Times of Covid-19. Online Brief. Belgrade. Social Inclusion and Poverty Statistical Office of the Republic of Serbia. 2021. Reduction Unit. Accessed October 15, 2021. Labor Force Survey database. Available at: Available at: http://socijalnoukljucivanje.gov.rs/ https://opendata.stat.gov.rs/odata/?id=en-us. en/social-safety-nets-in-times-of-the-covid- 19-crisis-2/ Statistical Office of the Republic of Serbia. 2020. Household budget Survey, Income and living Government of the Republic of Serbia. 2018. Third conditions Survey database. Available at: National Report on Social Inclusion. Belgrade. https://opendata.stat.gov.rs/odata/?id=en-us. Hirose, K., and Z. Czepulis-Rutkowska. 2016. Chal- UNICEF and World Bank. 2021. Serbia Human lenges in Long-Term Care of the Elderly in Development Review, forthcoming. Central and Eastern Europe. Budapest: Inter- national Labor Organization. UNICEF and World Bank. 2021b. Serbia: Social Assistance Thematic Note. Unpublished mimeo. Back to table of contents 25 SERBIA Social Protection Situational Analysis UNICEF. 2020. Study on the Effects of Covid-19 Pandemic on Families with Children in Serbia, June-July 2020. New York. Vienna Institute for International Economic Studies (WIIW) and the World Bank. 2020. Southeast Europe Jobs Gateway. Available at: https:// data.wiiw.ac.at/seejobsgateway-q.html World Bank. 2022. Social Protection Expenditure and Evaluation Database. Internal World Bank Database World Bank. 2021a. World Development Indicators. Available at: https://databank.worldbank.org/ source/world-development-indicators World Bank. 2021. Western Balkans Regular Economic Report No.19. Subdued Recovery. Spring 2021. Washington DC: World Bank Group. World Bank. 2019. Serbia Country Economic Memorandum: Serbia’s New Growth Agenda. Washington DC: World Bank Group. World Bank. 2018. Serbia: Efficiency Analysis of the National Employment Service – Update 2012- 2016. Unpublished. World Bank. 2017. Vertical Functional Review of Service Delivery in Social Services Sector, Synthesis Report. Unpublished. World Health Organization (WHO). 2014. WHO Global Nutrition Targets 2025: Stunting Policy Brief. Available at https://www.who.int/ nutrition/topics/globaltargets_stunting_poli- cybrief.pdf Back to table of contents 26 SERBIA Social Protection Situational Analysis Annex 1. Legal Framework The main legal framework guiding the social protection system includes the following key pieces of legisla- tion. The Social Welfare Law (2011), the Law on Financial Support to Families with Children (2017, amended in 2018), and the Law on Pension and Disability Insurance (2003, amended in 2019). Additional legislative pillars include the Law on Contributions for Mandatory Social Insurance (2004, 2012 setting contribu- tion rates), Decree on the Protection of Energy Vulnerable Customers (2014 and from 2015) and, for labor markets, the Law on Employment and Insurance in Case of Unemployment (2003, amended in 2017) and the Labor Law (2005, amended in 2017). • The Social Welfare Law (2011, latest bylaw 2016) is the main piece of legislation for social assistance and social services. The law establishes the objectives of improving protection of the poorest and increasing the efficiency of benefit systems and the development of community-level services. It regulates, inter alia, the financial social assistance and caregiver allowance. • The Law on Financial Support to Families with Children (2017) regulates the provision of financial support with both fertility and social protection objectives. This law covers cash benefits as a means of protecting poor families with children (child allowance) as well as benefits aimed at work-parenthood reconciliation and birth support (birth-related leave pay and parental allowance). With the new law, the parental allowance received a significant increase in the amounts allocated. • The Law on Pension and Disability Insurance (2003, latest amendment in 2019)67 regulates the compul- sory retirement and disability insurance. The law introduced the point structure in the system, one minimum pension, provided a new definition of disability, increased retirement ages and introduced rule-based indexation into the system. Key parameters have been substantially revised since 2003 including on: retirement age, indexation formula and value of minimum pensions. In addition, the government has introduced ad-hoc pension increments in certain years. Contribution rates, set by a specific Law, have also changed. • The Decree on Protection of Energy Vulnerable Customers provides for the energy deduction bill for energy vulnerable households (2014 and 2015 currently under revision). The Decree is currently being updated with the aim to increase the number of eligible energy vulnerable beneficiaries. • The Law on Employment and insurance in case of unemployment regulates, inter alia, the activities of the National Employment Service (NES) and the conditions for unemployment insurance and active labor market programs; the Labor Law dictates the condition of employment, including termination. Back to table of contents 27 A subsequent amendment was passed in 2021. 67 SERBIA Social Protection Situational Analysis Annex 2. Pension Parameters in Serbia Retirement age 65 for males/63 and 4 months for females (gradually raised by six months a year be- tween 2014 and 2020 and then by two months a year until reaching 65 years in 2032) with 15 years of contribution. Any age with 45 years of contribution. Early pension: Early pension: in 2020 - 58 years and four months for men / 57 years and 8 months for women (gradual increase to 60 units by 2023 for men, and by 2024 for women) with at least 40 years (men) / 39 years (women with a gradual increase to 40 years by 2023) contributions. The pension is reduced by 0.34 percent for each month it is claimed before the normal retirement age, up to 20.4 percent. Pension calculation The pension is calculated based on the number of years of contributions (up to 45 years), the ratio of the individual's gross earnings to the national average annual wage in each year of contributions starting from 1970, and the value of the general point. Valorization Value of the general point is determined each year and it is indexed using the same rule that applies to pensions in payment. Indexation post- Law approved in December 2019 to reintroduce Swiss formula indexation (according retirement to prices and wages). Eligibility for disability Contributed for 1 year if younger than 20; 2 years if 20-24; 3 if 25-29 and at least 5 pension years if age 30 years or older. No requirement if work injury (defined as inability to perform any work). Level of disability pension The pension is calculated based on the number of years of contributions, the ratio of the individual's gross earnings to the national average annual wage in each year of contributions starting from 1970, and the value of the general point. Eligibility for survivor’s The deceased was a pensioner or had at least five years of coverage. pension Eligible survivors include a widow age 53 years or older or a widower age 58 years or older who is disabled or caring for a child younger than 15 years (26 years if a student; no limit if disabled); a dependent mother age 60 years or older or disabled; a depend- ent father age 65 years or older or disabled; children younger than age 15 (age 26 if a student; no limit if disabled); and dependent grandchildren, brothers, and sisters. A widow(er) must have been married or have been in cohabitation with the deceased for at least three years or had a child with the deceased unless the deceased was age 65 years or older (men) or age 60 years or older (women) at the time of marriage. The widow(er)’s pension does not cease upon remarriage, except in the case of a spouse or extramarital partner of a professional military person according to the regulations on the Serbian Army who died during the action. Level of survivor’s 70 percent of the old-age pension the deceased received or was entitled to receive is pension paid for one survivor (140 percent for a full orphan); 80 percent for two survivors (160 percent for full orphans); 90 percent for three survivors (180 percent for full orphans); or 100 percent for four or more survivors (200 percent for full orphans). The pension is split equally among all eligible survivors. The minimum survivor pension is the old-age pension calculated based on 20 years of coverage. Contribution rates 25.5 percent of gross salary - 12 percent by employer and 14 percent by employee reduced from 26 percent to 25.5 percent by law approved in December 2019).68 Back to table of contents 28 The contribution rate for pension and disability insurance from January 1, 2022, was 25 percent (11 percent of salary - paid by the employer, and 14 percent of salary - paid by 68 the insured). The contribution rate was reduced in 2021, with amendments to the Law on Contributions for Compulsory Social Insurance. © 2022 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This project is funded by the European Union