A NOTE ON PERFORMANCE OF SOCIAL PROTECTION PROGRAMS IN CROATIA Claudia Rodriguez Alas, Michele Davide Zini and Aleksandra Posarac1 May 2019 Abstract Even though on a declining trend, the 2017 at-risk-of poverty or social exclusion rate (AROPE) in Croatia, at over 26 percent, was well above the EU-28 average of 22 percent. One of the most striking facts about the risk of poverty or social exclusion in Croatia is that it affects older people in particular. Among the population above 65, the share of those at risk of poverty and social exclusion in Croatia was almost twice as high as that recorded across the EU-28 (32.7 percent vs. 18.2 percent). As an EU member state, Croatia is a part of the Europe 2020 strategy and, accordingly, aims to reduce the number of persons living at risk of poverty or social exclusion, and inequality. The effectiveness of the country’s expenditure on social protection programs, and their design, play an important role towards the achievement of these goals. This note analyses how social protection programs perform in terms of expenditures, coverage, incidence, level of benefits and their impacts at reducing poverty and inequality. Croatia has a mature, well established social protection system that includes social insurance, social assistance and labor market programs. A significant portion of GDP is allocated to social policy (about 21 percent), but this figure is still below the EU-28 average of 28 percent. Two-thirds of these resources are devoted to contributory sickness and health care benefits (Croatia has a universal health insurance scheme), and old-age benefits. Another 10 percent is absorbed by contributory unemployment benefits. Thus, like in many other social protection systems in the European Union, in Croatia contributory programs dominate. Non-contributory programs, which are designed as complementary to contributory programs, represent a small portion of the overall spending on social protection. They are mostly comprised of universal categorical and, to a smaller extent, means-tested interventions. In terms of coverage, a large percentage of the total population participate in social protection programs. Social assistance benefits are well-targeted and are the most pro-poor among social protection benefits. Croatia’s social protection programs have a very large impact on poverty and inequality. On average, social protection beneficiaries are almost equally distributed among genders. Reflecting a diverse set of categorical programs, the age distribution of beneficiaries varies considerably between programs. However, only 6 percent of social assistance beneficiaries are above 65, which is at odds with the risk of poverty that this age group faces. As noted, Croatia has a well performing social protection system. However, some room for further development and improvements exist. First, Croatia should consider increasing the social assistance program’s coverage among the poorest 20 percent of the population. This should be done by paying attention to the demographic groups which face the highest risk of poverty (i.e. the over-65). Second, if politically feasible, the resources devoted to means-tested programs should be increased to achieve even larger impacts on poverty. 1The authors would like to acknowledge the kind support of the Croatian Bureau of Statistics in accessing the data used for the purposes of this analysis. Results and conclusions are of the authors alone and not those of the Croatian Bureau of Statistics. The team would like to thank Natalia Millan, Sandor Karacsony, Lars Sondergaard, Ana Simundza and Sabine Perrissin for comments and feedback, as well as Zoran Anusic and Ivan Drabek for inputs. 1 TABLE OF CONTENTS INTRODUCTION .............................................................................................................................................................3 I. WHAT ARE THE MAIN FEATURES OF POVERTY & SOCIAL EXCLUSION? ................................................................4 II. WHICH ARE THE MAIN SP PROGRAMS AND HOW MUCH RESOURCES ARE DEVOTED TO THEM? ......................9 Overview of Main Social Protection Programs in Croatia .........................................................................................9 Social Insurance ....................................................................................................................................................9 Social Assistance .................................................................................................................................................10 Expenditure on Social Protection Programs in Croatia ...........................................................................................11 III. HOW DO SOCIAL PROTECTION PROGRAMS’ PERFORM? ...............................................................................16 What Is the Coverage Of Social Protection Programs .............................................................................................17 Are benefits Pro-Poor? ............................................................................................................................................20 How Generous are Benefits?...................................................................................................................................22 What is the Impact of SP Programs on Poverty and Inequality? .............................................................................24 IV. WHAT IS THE PROFILE OF SOCIAL PROTECTION BENEFICIARIES? ..................................................................27 CONCLUSIONS .............................................................................................................................................................29 2 INTRODUCTION Croatia’s has a complex social protection system administered at the national, county and city levels. A review of the programs done by Šućur et al.2 documented about 130 social benefits managed by the national government and about 3,000 benefits managed by cities, municipalities and counties. With significant public resources allocated to social protection programs3 (21.3 percent of the GDP in 2016), it is crucial to assess the performance of such programs. Are social protection programs reaching their objectives? Are they cost-effective? What is the programs’ impact on reducing poverty and inequality? The answers to these questions can provide policy makers and program managers with much needed evidence to make informed decisions about social policy development. The objective of this note is to analyze the performance of Croatia’s social protection programs. There is no comprehensive information on the vast number of social protection programs implemented in the country due to the system’s complexity and fragmentation. However, this note uses the most reliable information available: Eurostat/ESPROSS database4 (European System of Integrated Social Protection Statistics), Croatia’s Household Budget Survey (HBS) and the World Bank’s ASPIRE database5. For the analysis of the social protection system’s performance, four rounds of the HBS are used (2008, 2010, 2011 and 2014) to present trends over time. Given that the HBS is nationally representative, performance of SP programs can be assessed at a national scale making it possible to identify some gaps in policies, for instance, population that may be eligible, but it is not covered by SP programs. The survey also provides a measure of household consumption which this note uses to assess distributional effects of the social protection transfers on the welfare of beneficiaries. Using administrative and household survey data, this note provides an overview of the level of spending and performance of social protection programs. The note is divided in five main sections: (1) A short review of the main facts on the risk of poverty and social exclusion in Croatia; (2) A description of Croatia’s main social protection programs and an analysis of expenditures on social protection in the country, breaking down the composition and trends of social expenditure using ESPROSS methodology; (3) An assessment of the performance of social protection programs, using four main indicators: coverage, benefits and beneficiary incidence, level of benefits and impact on poverty and inequality, and (4) A short profiling exercise of social protection beneficiaries, providing an overview of their gender, age, and labor market status using HBS data; (5) A final section concludes and presents some options on how to improve the country’s SP system. The performance analysis includes indicators by groups of benefits or “functions” and focuses on three “narrow” social protection areas: social insurance, social assistance, labor market policies. Due to data limitations, the performance analysis excludes health and sickness benefits, contributory and noncontributory. It also provides benchmarking of indicators by comparing Croatia with other European countries. 2 Šućur et al., 2016, cited in Urban et al., 2017. 3 These include contributory and non-contributory sickness and health care benefits, old age benefits, disability benefits, survivors’ benefits, family/children benefits, unemployment benefits, housing benefits, and social exclusion benefits not elsewhere classified, which are programs such as guaranteed minimum income schemes aimed at addressing social exclusions not covered by any other function. 4 ESPROSS is a common framework to harmonize social protection indicators from administrative national data. ESPROSS was developed by Eurostat and the European Union in the late 1970’s to provide comparable data between European countries. 5 The Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE), is a World Bank harmonization effort to provide comparable performance and social expenditure indicators between countries around the world. 3 I. WHAT ARE THE MAIN FEATURES OF POVERTY & SOCIAL EXCLUSION? Even though on a declining trend, the 2017 at-risk-of poverty or social exclusion rate (AROPE)6 in Croatia, at over 26 percent, was well above the EU-28 average of 22 percent. As the effect of the global financial crises unraveled, the AROPE rate in Croatia rose, peaking at over 32 percent in 2011 and 2012, almost 8 percentage points higher than the EU28 average. The trend began to reverse after 2012, and the AROPE rate has been declining steadily since, settling at 26 percent in 2017, 4 percentage points above the EU figure, but still in the upper half of the EU’s distribution. Figure 1. At-risk-of poverty or social exclusion rate, 2010-2017, Croatia vs EU28 (Share of population) 34 32 30 28 26 26.4 24 22 22.4 20 2010 2011 2012 2013 2014 2015 2016 2017 European Union - 28 countries Croatia Figure 2. At-risk-of poverty or social exclusion rate, 2017 (Share of population) 50 40 30 20 10 0 Czechia EU-28 Malta Denmark Croatia Latvia Portugal Germany Austria Sweden Greece Spain Cyprus Slovenia Luxembourg Poland Slovakia Romania Lithuania Switzerland (²) Norway (²) Hungary Belgium France Italy Ireland Netherlands Finland Iceland (²) Turkey (²) Bulgaria Serbia Estonia United Kingdom North Macedonia (²) Very low work intensity (other than at risk of poverty or experiencing severe material deprivation) Severe material deprivation (other than at risk of poverty) At risk of poverty after transfers (income poverty) Source: Eurostat https://ec.europa.eu/eurostat/web/social-protection/data/main-tables 6According to EUROSTAT, the “at risk of poverty or social exclusion, abbreviated as AROPE, corresponds to the sum of persons who are either at risk of poverty, or severely materially deprived or living in a household with a very low work intensity (divided by the total population, right?). Persons are only counted once even if they are present in several sub-indicators. The AROPE rate, the share of the total population which is at risk of poverty or social exclusion, is the headline indicator to monitor the EU 2020 Strategy poverty target.” See https://ec.europa.eu/eurostat/statistics- explained/index.php?title=People_at_risk_of_poverty_or_social_exclusion 4 The risk of poverty or social exclusion in Croatia affects older people in particular. As figure 3 below shows, the AROPE rate in 2017 among children, at 25.8 percent was comparable to that in the EU28 (24.9 percent). Among adults between the age of 18 and 64, the AROPE rate in Croatia was above the EU average, but not dramatically so (24.6 percent vs. 23 percent, respectively). However, once the analysis is restricted to those above 65, the share of the population at risk of poverty and social exclusion in Croatia was almost twice as high as that recorded across the EU-28 (32.7 percent vs. 18.2 percent). Figure 3. At-risk-of poverty or social exclusion rate, by age group, 2017 (Share of population) 35 32.7 30 24.9 25.8 24.6 25 23.0 18.2 20 15 10 5 0 Children Adults Older people (aged 0-17 years) (aged 18-64 years) (65 years and over) EU-28 Croatia Source: Croatian Bureau of Statistics The risk of poverty is somewhat higher among women and, not surprisingly, affects the unemployed, and those with lower levels of education relatively more than those with higher levels of education. Figure 4 shows the risk of poverty rate (i.e. the level of income poverty7) disaggregated by gender. In 2017, the risk of poverty was 18.9 percent among men, but 20.9 percent among women. As per figure 5, among those with basic education levels, the risk of poverty reaches almost 40 percent, but drops to about 4 percent among those with tertiary education. Figure 4. At-risk-of poverty or social exclusion rate, by gender, 2017 (Share of population) 23 22 21 21 21 21 21 20 20 20 21 20 21 20 20 20 20 20 20 19 19 19 19 19 19 19 19 17 15 13 11 9 7 5 2010 2011 2012 2013 2014 2015 2016 2017 Total Men Women Source: Croatian Bureau of Statistics 7The at-risk-of-poverty rate is the share of people with an equivalized disposable income (after social transfer) below the at-risk- of-poverty threshold, which is set at 60 % of the national median equivalized disposable income after social transfers. 5 Figure 5. At-risk-of poverty or social exclusion rate, by level of education, 2017 (Share of population) 40 37 37 36 36 37 36 36 35 35 30 25 20 16 16 16 16 16 16 16 15 15 10 6 6 6 6 5 5 5 4 5 0 2010 2011 2012 2013 2014 2015 2016 2017 Preschool and basic Upper secondary Tertiary Source: Croatian Bureau of Statistics The risk of poverty increases dramatically among those not at work (unemployed, retired and other inactive), relative to those at work (employees and self-employees). As figure 6 shows, the risk of poverty is on average 5.8 percent among the employed and self-employed but increases to over 30 percent among those out of work. Among those at work, the self-employed fare worse than employees by over 10 percentage points (15.6 percent vs. 4.8 percent). Among those out of work, the unemployed, especially women, fare particularly bad in terms of being at risk of poverty, with 45 percent of the jobless being at risk of poverty (52 percent among unemployed women alone). Figure 6. At-risk-of poverty or social exclusion rate, by labor market status, 2017 (Share of households) 60 52 50 46 40 40 33 31 31 30 31 30 27 25 24 22 20 16 15 16 10 6 5 7 5 4 6 0 At work Not at work Employees Self- Unemployed Retired Other employed inactive Total Men Women Source: Croatian Bureau of Statistics 6 The risk of poverty is especially high among single member households and, as mentioned above, worsens if we consider those older than 65, living alone. While the risk of poverty among all households without children was 22 percent in 2017, it increased to 44 percent among single member households (almost twice as much as within the EU28), and to 48 percent among people above 65 living alone. Figure 7. At-risk-of poverty or social exclusion rate, by household type, 2017 (Share of households) – Households without Dependent Children 60 50 40 30 22 20 10 0 All One-person One-person One-person Two adults Two adults, Two adults, Two or more Three or Households household household, household, both under at least one adults more adults without person under 65 years or 65 adult 65 dependent 65 years over years years or over children Source: Source: Croatian Bureau of Statistics Among households with dependent children, single parents, and households with three or more children, face a high risk of poverty. While the risk of poverty among all households with children was 18 percent in 2017, this risk increased to 31 percent for households with 2 adults and 3 or more children, and to 37 percent for single parent households. Figure 8. At-risk-of poverty or social exclusion rate, by household type, 2017 (Share of households) – Households with Dependent Children 40 35 30 25 20 18 15 10 5 0 All Households Single parent Two adults with Two adults with Two adults with Two or more Three or more with dependent with one or one dependent two dependent three or more adults with adults with children more dependent child children dependent dependent dependent children children children children Source: Source: Croatian Bureau of Statistics 7 Inequality indicators have improved in recent years and are aligned with EU figures. In 2017, the top 20 percent of the population (with the highest income) received 5.1 times as much income as the bottom 20 percent, down from 5.5 in 2010 and in line with the EU-28 average of 5.0 in 2017.8 Similarly, the Gini coefficient in Croatia declined from 31.6 percent in 2010 to 29.9 percent in 2017, below the EU-28 average of 30.7 percent. Figure 9. Inequality Indicators in Croatia 32 5.8 32 5.6 31 5.4 31 5.2 30 5 30 29 4.8 29 4.6 2010 2011 2012 2013 2014 2015 2016 2017 Gini coefficient Income quintile share ratio (S80/S20) Source: Eurostat https://ec.europa.eu/eurostat/web/social-protection/data/main-tables As an EU member state, Croatia is a part of the Europe 2020 strategy and, accordingly, aims to reduce the number of persons living at risk of poverty or social exclusion, and inequality within the country. Achieving that goal depends crucially on the effectiveness of the country’s expenditure on social protection programs, and on the design of solid policy tools. In the next sections we describe the main programs of Croatia’s social protection system analyze their performance. 8The quintile share ratio is the proportion of the total equivalized disposable income received by the 20% of the population with the highest income (top quintile) to that received by the 20% of the population with the lowest income (lowest quintile). 8 II. WHICH ARE THE MAIN SP PROGRAMS AND HOW MUCH RESOURCES ARE DEVOTED TO THEM? Overview of Main Social Protection Programs in Croatia9 Croatia has a wide list of programs to help households and individuals – particularly poor and vulnerable ones – cope with crises and shocks, find jobs, invest in their own health and education and that of their children, and protect households from income risks related to aging. There are about 130 social protection benefits managed by the national government and about 3,000 benefits managed by cities, municipalities and counties. These programs are either contributory or noncontributory, qualifying as social insurance or social assistance tools, respectively. Below we list the main programs managed by the national government. Active labor market programs such as on-the-job trainings, internships, skilling and reskilling programs, lifelong trainings for adults or the unemployed, etcetera, are not included in the list below, but are nevertheless included in the analysis of expenditures in the sections that follow. Social Insurance Social insurance programs include: i) old age, disability and survivor benefits; ii) sickness/health benefits, and work injury benefits; and iii) unemployment benefits. i) The main nation-wide old age, disability and survivor programs include: o An old-age pension, accessible by men after the age of 65 and by women after the age of 62 (although the latter has been rising by three months a year until it will reach age 65 in 2030). Both men and women need to have been covered by social insurance for at 15 years. The retirement age for both men and women will gradually rise by three months a year from 2031 until reaching age 67 in 2038. o A general disability pension, which is granted following the total and permanent loss of ability to work due to health reasons occurring before age 65. The insured must have had social insurance coverage for at least one-third of his/her working life after age 20 (or age 23 for insured persons with postsecondary education; and age 26 for a university graduate). There is no minimum qualifying period if the disability is the result of a work injury or an occupational disease. o A partial disability pension for those younger than 65. If the disability began before age 53, the insured must have a reduced capacity for work that is not likely to return to full capacity with occupational rehabilitation. However, the person should be able to work at least 70% of daily working hours in a different job requiring the same level of education as the job before the disability began. The insured must have had social insurance coverage for at least one-third of his/her working life. There is no qualifying period if the disability is the result of a work injury or occupational disease. o An occupational rehabilitation and salary compensation is granted if the disability began before age 53 and the insured is likely to regain the capacity to work full time (40 hours a week). The insured is eligible for salary compensation during rehabilitation, until s/he returns to work in another job with the same employer. If no other job is available or suitable, the insured will receive the compensation for up to 12 months of unemployment following completion of occupational rehabilitation or for 24 months if the disability is the result of a work injury or an occupational disease. o A temporary disability pension is paid to disabled persons following occupational rehabilitation who remain unemployed for at least five years prior to reaching age 58. o A compensation allowance for a physical injury is paid for the loss of or damage to a part of the body or an organ because of a work injury. The injury must be assessed at 30% or more. 9This section is based on: https://www.ssa.gov/policy/docs/progdesc/ssptw/2018-2019/europe/croatia.html and “Assessment of Social Benefit Effectiveness: An analysis at the municipal, and city level of Croatia”, a Project co-financed by the European Union from the European Fund for Regional Development, September 2017. 9 o A disability pension is accessible to those aged 55 or older who meet the qualifying conditions for the social insurance disability pension and are members of a compulsory pension fund for 10 years. o A survivor pension is extended to pre-defined survivors of an active worker, deceased pensioner or an occupational rehabilitation beneficiary who had five years of contributions, 10 years of coverage, or met the qualifying period conditions for a social insurance disability pension. There is no minimum qualifying period if the insured's death was the result of a work injury or an occupational disease. ii) The main sickness/health benefits include: o A cash sickness benefit for those with at least nine consecutive months of contributions or at least 12 nonconsecutive months of contributions in the two years before the incapacity began. If the insured does not meet the contribution requirements, a minimum benefit may be paid. o A cash maternity and parental leave benefits, with the same eligibility requirements of the cash sickness benefit. o A one-off newborn child assistance paid for the birth of a child to those also with at least nine consecutive months of contributions or at least 12 nonconsecutive months of contributions in the two years before the birth. o Work injury benefits can be granted in the form of cash to workers covered by social insurance following a work injury or an occupational disease as determined by a designated doctor in a primary health care institution, and for the period of incapacity. iii) Unemployment benefits Unemployment benefits are accessible by employed workers with an employment contract aged 15 to 65 with at least nine months of employment or self-employment in the 24 months prior to the fall in unemployment. Beneficiaries must be involuntarily unemployed; capable of, available for, and actively seeking work; and must register with the Employment Service in the first 30 days of unemployment. Social Assistance The main social assistance program at the central level include: o The Guaranteed Minimum Benefit (GMB) is the country’s core social assistance program of last resort for the poorest households. The amount of the income and means-tested benefit, intended for families and individuals to meet their basic needs, depends on household considerations such as: the number of adults and children, work capability of members, and whether there is a single parent. o Family allowances are extended to parents or guardians who have resided in Croatia for at least three years who meet certain income criteria o An income and means tested allowance for nursing and care is extended to persons who cannot meet the life’s basic needs alone but need the assistance and care of another person in the organization of nutrition, preparation and consumption of meals, purchasing food, cleaning their living quarters, dressing, maintaining personal hygiene, and alike. o A means-tested personal disability allowance is given to a person with severe disability or other severe and permanently impaired health status to meet her/his needs for inclusion in the life of the community. o The parent caring for children program provides a monthly allowance, pension and health insurance coverage to the parents of a child with developmental disabilities or of a disabled person who is either completely dependent on the assistance and care of another. o One-Off Allowance is approved in cases of sudden difficulties in meeting basic life needs arising from the birth or education of a child, illness or death of a family member, natural disaster and the like. The allowance can be given in cash or in kind as a one-off payment without restriction on the number of approvals within the year, but its total amount is capped per year per household. 10 Expenditure on Social Protection Programs in Croatia10 From 2008 to 2017, Croatia spent a yearly average of 71.5 billion Kuna on social protection, or 21 percent of the country’s GDP. Expenditure on social protection programs reached 76.3 billion Kuna in 2016, of which over 75 billion (about 98 percent of total budget) were absorbed by benefits (i.e. non- administrative costs) 11. Figure 10, panel B, shows that social expenditure as a share of the GDP has remained roughly constant in the last years; however, the share reported in 2017 (at just over 20 percent) was almost two percentage points higher than in 2008, due mostly to an increase in the budgetary allocation to family/children and old age benefits. Figure 10. Croatia’s Total Social Protection Expenditure Panel A. Total Social Protection Expenditure in Million Kuna 100,000 80,000 60,000 40,000 20,000 0 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sickness/Health care Disability Old age Survivors Family/Children Unemployment Housing Social exclusion not elsewhere classified Administration costs Panel B. Total Social Protection Expenditure (as a share of GDP) 25 20 15 6.2 6.4 6.5 6.6 6.9 7.2 7.1 7.1 7.1 5.7 10 5 6.2 7.0 7.0 7.0 7.4 6.9 6.8 7.1 7.0 6.9 0 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sickness/Health care Disability Old age Survivors Family/Children Unemployment Housing Social exclusion not elsewhere classified Source: Croatian Bureau of Statistics – Classifications aligned with ESPROS Methodology 10 To analyzeCroatia’s social expenditure composition and trends this section uses data from the Croatian Bureau of Statistics, as well as ESPROSS and ASPIRE data. ESPROSS methodology groups social protection data in eight functions, based on the primary purpose of the social protection program. These functions are: old age, disability, survivors, family/children, unemployment, housing, and social exclusion not elsewhere classified, which are programs to address social exclusions not covered by any other function. Unlike ASPIRE, ESPROSS does not make the distinction between contributory schemes (social insurance) and non-contributory programs that are entirely funded by government resources (social assistance). 11 Croatian Bureau of Statistics, 2019. 11 Croatia’s expenditure on social protection is significant, but still lower than the European average.12 Figure 11 shows that in 2016 Croatia’s spent 20.9 percent of its GDP on social protection, which is 6 percentage points below the average spending in the European Union (28 countries). France reports the highest spending as percentage of the GDP (34 percent) while Romania reports the lowest (15 percent). Croatia spends comparatively more than most of the Central and Eastern European members of the EU. As in all European countries, non means-tested programs, comprise most of the social spending. In Croatia, about 19.9 percent of the GDP is devoted to non means-tested programs while only 1 percent is spent on means-tested programs – almost 4 and 2 percent lower than the EU-28 average, respectively. Figure 11. Total Social Protection Expenditure of European Countries in 2016 – Percent of GDP 35 30 27.1 25 20.9 20 15 10 5 0 EU-28 Austria Denmark Serbia Latvia Turkey France Finland Italy Norway Greece Hungary Romania Netherlands Slovenia Croatia Poland Cyprus Spain Bulgaria Ireland Belgium Luxembourg Slovakia Estonia Malta Lithuania Sweden Switzerland Portugal Czechia United Kingdom North Macedonia Germany Non means-tested benefits Means-tested benefits Source: Eurostat https://ec.europa.eu/eurostat/web/social-protection/data/main-tables Croatia’s expenditure on the administration of SP programs has been decreasing over time and it is now among the lowest in Europe.13 Figure 12 shows how the administrative costs of social protection program has decreased over the years, from 2.6 of GDP in 2008 to 1.6 percent in 2016. Compared to the EU-28 average of 2.7 percent of GDP, Croatia’s share is among some of the lowest in the continent according to Eurostat. Macedonia has the lowest administrative cost of social protection (0.2 percent of GDP) while Switzerland has the highest cost (6.1 percent of the GDP). 12 Expenditure on social protection contains: social benefits, which consist of transfers, in cash or in kind, to households and individuals to relieve them of the burden of a defined set of risks or needs; administration costs; other expenditure, which consists of miscellaneous expenditure by social protection schemes (payment of property income and other) (source: Eurostat). 13 Administrative costs represent the costs charged to the scheme for its management and administration (source: Eurostat). 12 Figure 12. Administrative Cost of Social Protection Programs Panel A. Administrative Costs as a % of GDP - Croatia 3 2.59 2.5 2.27 2.14 1.99 1.86 1.90 2 1.72 1.62 1.61 1.5 1 0.5 0 2008 2009 2010 2011 2012 2013 2014 2015 2016 Panel B. Administrative Cost of Social Protection Programs of European Countries in 2016 – Percent of GDP 6 5 4 3 2.7 2 1.6 1 0 Turkey Serbia Spain (p) Norway Malta (es) Romania Finland Austria Czechia (es) France Denmark Cyprus Ireland Latvia (p) Poland Netherlands EU-28 (p) Luxembourg Hungary (p) Croatia Estonia Sweden (p) Germany (p) Belgium Slovakia (p) Switzerland (p) Portugal Slovenia (p) Italy (p) North Macedonia (p) Greece (p) Bulgaria (es) Lithuania (p) United Kingdom (p) Source: Eurostat https://ec.europa.eu/eurostat/web/social-protection/data/main-tables (p) Provisional value; (es) Estimated value Contributory social protection programs comprise most of the spending. Two-thirds of the total budget is allocated to sickness and health care benefits, and old-age benefits; and another 10 percent is absorbed by contributory unemployment benefits. Old age benefits – which are almost exclusively contributory - constitute 35 percent of total spending, while sickness and health care benefits account for another 34 percent. The third larger set of benefits are disability related benefits, which account for 11 percent of total social protection expenditure and are 90 percent contributory. Family and children benefits, which absorb almost 9 percent of the total social protection budget, are almost universal (only 21 percent of these benefits is means tested). Social exclusion benefits not elsewhere classified (i.e. the Guaranteed Minimum Benefit), and housing benefits represent the smallest fraction of total spending (1.3 and 0.1 percent respectively; 4.5 percent of the total spending on social protection). Conceptualized as a last resort benefits and complementary to contributory programs, these benefits are means tested. 13 Figure 13. Expenditure by Social Protection Function – Percentage of Total Spending (2017) Unemployment, 2.1 Housing, 0.1 Social exclusion not Family/Children, 8.6 elsewhere classified, 1.3 Survivors, 8.9 Sickness/Health care, 33.7 Old age, 34.6 Disability, 10.7 Source: Croatian Bureau of Statistics Old age pensions, the largest social protection spending category in Croatia, is below the European average and it is almost exclusively non-means tested. In 2016, spending in old age pensions absorbed 7.1 percent of Croatian GDP (with only 0.1 percent of this being means-tested). Meanwhile the average spending in the EU-28 on old age pensions was 4 percentage points of GDP higher (10.9 percent), and about 0.5 percent of GDP in the average EU-28 country was devoted to means-tested old age programs (see figure 14). Figure 14. Old age pensions – Percentage of Total Spending (2017) 16 14 10.9 12 10 8 7.1 6 4 2 0 EU-28 Serbia Turkey Italy France Finland Austria Norway Romania Latvia Netherlands Hungary Greece Spain Poland Germany Cyprus Bulgaria Croatia Ireland Switzerland Slovenia Malta Slovakia Portugal Luxembourg Belgium Sweden Czechia Estonia Lithuania United Kingdom North Macedonia Non means-tested benefits Means-tested benefits Source: Eurostat https://ec.europa.eu/eurostat/web/social-protection/data/main-tables 14 Although social assistance programs represent a small portion of the overall spending on social protection, their share in GDP is significant. In targeting social assistance, Croatia mostly relies on categorical targeting, with only limited use of means-testing. Total spending on social assistance was 3.36 percent of GDP in 2014. The largest share went to veterans’ pensions, which account for almost half of the spending on social assistance programs (see figure 15). Child and family benefits account for another 0.80 percent of GDP. Means-tested program account for less than 1 percent of GDP (and include the last resort social assistance – i.e. the Guaranteed Minimum Benefit, and programs included under the “other social safety net programs, which include in-kind social programs). Figure 15. Social Assistance Expenditure – Percent of GDP War veteran pensions (2012) 1.47% Child and Family Benefits (2014) 0.80% Other social safety net programs (2014) 0.62% Last-Resort Social Assistance (2011) 0.30% Personal disability allowance (2014) 0.10% War-disabled veterans and civilians (2014) 0.05% Child care (both cash and in-kind) (2014) 0.005% Refugees and IDP (2014) 0.004% Source: World Bank’s Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) Notes: In parenthesis is the year of data used for each program. 15 III. HOW DO SOCIAL PROTECTION PROGRAMS’ PERFORM? This section analyses the performance of Croatia’s social protection system using primarily the most recent household budget survey data available to the World Bank team (HBS 2014) . The analysis conducted relies on ASPIRE’s methodology and classification. ASPIRE classifies programs in three areas of social protection: 1) social insurance, which are contributory programs designed to protect individual and households against loss of income due to old age, disability or death. 2) Social assistance, which refers to non-contributory programs aimed to protect individuals and households from poverty and destitution. 3) Labor market programs which can be either contributory or non-contributory to protect individuals from the risk of unemployment or help them acquire skills or access employment opportunities. The analysis generates four main sets of performance indicators: program’s coverage, benefit/beneficiary incidence, benefits’ level, and impact on poverty and inequality. These indicators are generated for each of the main social protection programs included in the HBS, and at a more aggregated level for the three social protection areas listed above. Time trends are also presented through the analysis of three previous rounds of the HBS survey (2008, 2010, 2011). Finally, indicators for Croatia are benchmarked against other European countries. The performance analysis is conducted for nine out of ten social protection programs included in the HBS (see table 1). Redundancy payments were not included in the individual program analysis, because the number of observations is not representative enough to generate meaningful statistical results. A few caveats are important to note before presenting the results: • The grouping of programs used in the HSB does not exactly coincide with the grouping created by ESPROSS using administrative data. In this sense, performance indicators cannot be exactly linked to program expenditures on all social protection functions. • Performance analysis is inevitably biased towards larger and mostly cash-based benefits captured in the survey. HBS data does not capture small programs, nor social protection services including sickness and health benefits or active labor market programs. Despite these caveats, household survey data, and its analysis using ASPIRE methodology, provides a good estimation of the performance of Croatia’s main programs, and it is currently the only source of information which allows us to identify gaps in program coverage, as well as distributional impacts of programs. Table 1. Social Protection Programs included in Croatia’s Household Budget Survey SP Areas Social Protection Programs Old-age pension with all allowances Family pensions with all allowances Disability pensions with all allowances Redundancy payment at time of regular retirement (standard or early retirement) Social Insurance Redundancy payment on account of termination of employment due to permanent excess number of employees. Earning for sick leave exceeding more than 42 days, compensation for physical injury, nursing by another person, employment of disabled persons. Labor Market Unemployment and re-training benefits Family benefits (child allowance, maternity leave, newborn package, childcare) Social welfare assistance in cash Social Assistance Social welfare assistance in kind (food, fuel, clothing, etc.) Scholarships and awards received at school 16 What Is the Coverage of Social Protection Programs14 A large percentage of the total population participates in social protection programs. According to HBS2014 data, 71 percent of Croatia’s population benefits from at least one social protection program. The programs driving upwards this figure are old-age pensions and allowances (34 percent of the population), family benefits, which include child allowance, maternity leave, newborn assistance, and childcare (22 percent of the population), and disability pensions and allowances (13 percent). Most importantly, social protection programs achieve almost universal coverage of individuals in the poorest 20 percent with 96 percent of this population benefiting from at least one intervention. The largest interventions covering the poorest consumption quintile are still old age pensions (43 percent), family benefits (41 percent) and disability pensions (21 percent). Cash-based welfare programs reach only about 15 percent of the poorest quintile (see figure 16). Among the smallest programs in terms of population coverage are scholarships and in-kind social welfare programs, which cover 2 and 2.4 percent of the total population, respectively, and 2 and percent of the poorest quintile of per capita consumption. According to the Venn diagram below (see figure 17), 29 percent of the population does not participate in any SP program, about 58 percent of the population benefits from only one type of program (social insurance, social assistance or labor market), while about 13.6 percent receives more than one program from any of the three social protection areas. Figure 16. Share of Population Covered by Social Protection Programs as Captured in the 2014 HBS All social protection Old-age pensions Family benefits (child allowance, maternity, layette) Disability pensions Family pensions Unemployment and retraining benefits Social welfare in cash Sick leave benefits Social welfare in-kind Scholarships 0 10 20 30 40 50 60 70 80 90 100 Poorest quintile Total population Source: Croatia's Household Budget Survey 2014 Notes: Coverage is the proportion of the population that receives the transfer. Results are projected to the total population using the survey’s household expansion factor multiplied by the household size. The poorest quintile is calculated using per capita pre-transfer consumption. The indicator includes direct and indirect beneficiaries (all the members of a household were at least one individual receives a transfer). 14To assess coverage, all members living with the direct beneficiary of a given program are counted as beneficiary of said program. 17 Figure 17. Overlap of Program by Social Protection Area Source: Croatia's Household Budget Survey 2014 As a share of the total population, coverage increased from 67 to 71 percent between 2008 and 2010and has remained stable at this level since then. Contributory or social insurance programs make up the core of social protection programs with 53 percent of the total population benefiting from them, a figure essentially unchanged relative to 2008. Coverage of social assistance programs, at 28 percent of the total population in 2014, increased by 7 percentage points between 2011 and 2014, but was only 3 percentage points higher than the figure recorded in 2008. Among the poorest quintile, the system has experienced an upward trend in terms of coverage, increasing from 87 percent in 2008 to 96 percent in 2014, mostly due to social insurance programs. Coverage of social insurance programs has experienced a slow decrease since 2010 but, at 68 percent in 2014, still displays a sizeable increase in coverage relative to 2008 level of 59 percent. Social assistance programs, however, displayed in 2014 a level of coverage among the poorest quintile essentially unchanged relative to the 2008 level. Figure 18. Percentage of Population Covered by Social Protection Programs (2008-2014) A. Total Population 80 71 71 71 70 67 58 56 60 52 53 50 40 25 25 28 30 21 20 10 4 6 5 5 0 All social protection Social insurance Labor market programs Social assistance 2008 2010 2011 2014 18 150 B. Poorest Quintile 95 96 100 87 94 74 71 68 59 53 46 43 52 50 4 8 7 6 0 All social protection Social insurance Labor market programs Social assistance 2008 2010 2011 2014 Source: Croatia's Household Budget Survey 2008, 2010, 2011, 2014. Notes: Coverage is the proportion of the population that receives the transfer. Results are projected to the total population using the survey’s household expansion factor multiplied by the household size. The poorest quintile is calculated using per capita pre-transfer consumption. The indicator includes direct and indirect beneficiaries (all members of a household were at least one individual receives a transfer). Croatia’s coverage of social protection programs is among the highest among the selected European countries.15 Figure 19 below shows that only Romania and Poland display a slightly higher percentage of the poorest quintile covered by these programs (98 and 97 percent respectively). In terms of the total population, Croatia is the third highest after Romania and Russian Federation (84 and 78 percent respectively). Croatia’s social protection is above the average of 63 percent among the countries in Central and Eastern Europe, and Central Asia included in our dataset when it comes to coverage of the total population and the poorest quintile (average of 85 percent). Figure 19. Percentage of Total Population and Poorest Quintile Covered by Social Protection in Selected European Countries 100 95.6 90 80 71.1 70 60 50 40 30 20 10 0 Poorest Quintile Total Population Source: World Bank’s Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) Notes: Coverage is the proportion of the population that receives the transfer. Results are projected to the total population using the survey’s household expansion factor multiplied by the household size. The poorest quintile is calculated using per capita pre-transfer consumption. The indicator includes direct and indirect beneficiaries (all the members of a household were at least one individual receives a transfer). Not all high-income countries are included in the ASPIRE database, thus sample is limited to the data available. 15ASPIRE does not include data for most of the EU countries, so figures in this note only includes countries for which data/most recent data are available. 19 Are benefits Pro-Poor? The analysis of the incidence of benefits provides an assessment of how benefits are distributed among program participants. For example, a program is considered pro-poor if most of its benefits are distributed among the poorest. The incidence of benefits among different groups of beneficiaries depends on various factors, including program design (i.e. what the targeted population is) or issues with program delivery (whether, for instance, there are targeting errors leading to benefits leaking to non-intended beneficiaries). Figure 20 shows how Croatia’s SP benefits are distributed among quintiles of the welfare distribution. Social welfare benefits are the most pro-poor, as eligibility for such programs are mostly income-based. About 70 percent of social welfare benefits in cash and 65 percent of social welfare benefits in kind are received by the poorest quintile. A large percentage of family benefits are also transferred to the poor (41 percent). Contributory pensions, on the other hand, are less pro-poor with only 20 to 26 percent of the benefits going to the poorest quintile. This is to be expected, as social insurance benefits are not targeted based on income but are reserved for those who have contributed to the social insurance scheme. Figure 20 shows that old-age pensions are more equality distributed among each quintile of the welfare distribution than other social security benefits. Scholarships are least progressive since the incidence of benefits is higher for the richest quintile (46 percent) compared to any other welfare group. Figure 20. Distribution of SP Benefits, by Quintiles of Pre-Transfer Per Capita Consumption Social welfare in cash Social welfare in-kind Family benefits (child allowance, maternity, layette) Disability pensions Unemployment and retraining benefits Old-age pensions Sick leave benefits Family pensions Scholarships 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Poorest quintile Q2 Q3 Q4 Richest quintile Source: Croatia's Household Budget Survey 2014. Notes: Benefits incidence is the transfer amount received by a group as a percent of total transfers received by the population. Results are projected to the total population using the survey’s household expansion factor multiplied by the household size. The poorest quintile is calculated using per capita pre-transfer consumption. The share of social insurance benefits received by the poorest quintile increased between 2008 and 2014; however, the share of social assistance benefits reaching the bottom quintile decreased. The proportion of social assistance benefits going to the poorest 20 percent decreased from 54 percent in 2008 to 47 percent in 2014. In the case of social insurance, the proportion of benefits received by this 20 group increased 9 percentage points from 2008 to 2010 and then has remained stable at around 24 percent in the following years. Unemployment benefits experienced an even higher increase (13 percentage points) from 2008 to 2010; however, their incidence among the poorest 20 percent decreased since then - from 28 percent in 2010 to 23 percent in 2014. Even though in Croatia social assistance programs are mostly pro-poor, the share of benefits going to the poorest quintile is slightly below the average of selected countries. On average, among the 15 countries featured in Figure 22, about 49 percent of social assistance benefits go to the poorest quintile. Croatia’s incidence of benefits going to this group is two percentage points lower (47 percent). The country with the most pro-poor distribution of social assistance benefits is Kosovo (72 percent) while the lowest is Russian Federation with only about 30 percent of benefits accruing to the poorest 20 percent (Russia’s social assistance system features only few income/ means tested benefits). Figure 21. Percentage of Total Social Protection Benefits Received by the Poorest Quintile (2008-2014) 60 54 53 48 47 50 40 28 26 30 24 23 24 23 20 15 15 10 0 All social insurance All labor market programs All social assistance 2008 2010 2011 2014 Source: Croatia's Household Budget Survey 2008, 2010, 2011, 2014. Notes: Benefits incidence is the transfer amount received by a group as a percent of total transfers received by the population. Results are projected to the total population using the survey’s household expansion factor multiplied by the household size. The poorest quintile is calculated using per capita pre-transfer consumption. Figure 22. Percentage of Total Social Assistance Benefits Received by the Poorest Quintile – Selected Countries 80 70 60 47.1 49.0 50 40 30 20 10 0 Source: World Bank’s Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) Notes: Benefits incidence is the transfer amount received by a group as a percent of total transfers received by the population. Results are projected to the total population using the survey’s household expansion factor mu ltiplied by the household size. The poorest quintile is calculated using per capita pre-transfer consumption. Not all high- income countries are included in the ASPIRE database, thus sample is limited to the data available. 21 How Generous are Benefits? The value of contributory benefits (old-age, disability and family pensions and allowances) as a share of beneficiaries’ post-transfer consumption is the highest among all social protection programs. This is expected because social insurance programs are designed to replace working earnings and thus, the benefit levels are much higher than those of social assistance programs. Old-age pension for example, represent 48 percent of household post-transfer consumption for the total population and 43 percent for the poorest quintile. Cash-based social assistance programs also make up a substantial proportion of beneficiary households’ consumption (22 percent for the total population and 27 percent for the poorest quintile). The generosity of other programs is on the other hand relatively low, since they represent less than 20 percent of household total consumption. Evidence from impact evaluations from around the world suggests that the most effective cash transfer programs in terms of poverty mitigation contribute at least 20 percent to the beneficiary household consumption.16 Figure 23. Benefits as Share of Household Post-Transfer Consumption Old-age pensions Disability pensions Family pensions Social welfare in cash Sick leave benefits Family benefits (child allowance, maternity, layette) Unemployment and retraining benefits Scholarships Social welfare in-kind 0 10 20 30 40 50 60 70 80 90 100 Poorest Quintile Total Population Source: Croatia's Household Budget Survey 2014. Notes: Results are projected to the total population using the survey’s household expansion factor multiplied by the household size. For the overall population, the value of the social protection benefits as a share of beneficiary household consumption increased from 2008 to 2014 (see Figure 24, panel A). The value of benefits went from 33 percent of household consumption in 2008 to 44 percent in 2011, only to then marginally decrease to 42 percent by 2014. The upwards trend between 2008 and 2014 can be observed in all three areas of social protection. However, among the poorest quintile alone, the trend is not as positive. In 2014, social protection benefits accounted for less total household consumption relative to 2008 (from 46 to 43 percent), with social assistance programs being the only ones displaying a positive change in terms of transfers as a share 16 Handa el al. 2013. 22 of beneficiary consumption (from 16 to 20 percent between 2008 and 2014). The small decrease observed between 2011 and 2014 for social assitance programs may be a reflection of the introduction of the Guaranteed Minimum Benefit (GMB) in 2013, under which several social benefits were consolidated.17 Figure 24. SP Benefits as Percentage of Household Post-transfer Consumption (2008-2014) A. Total Population 50 48 47 44 43 45 42 40 40 36 35 33 30 25 20 15 13 13 13 11 11 9 8 9 10 5 0 All social protection Social insurance Labor market programs Social assistance 2008 2010 2011 2014 B. Poorest Quintile 60 54 50 51 50 48 48 46 45 43 40 30 21 22 20 20 16 16 16 16 12 10 0 All social protection & labor Social insurance Labor market programs Social assistance 2008 2010 2011 2014 Source: Croatia's Household Budget Survey 2008, 2010, 2011, 2014. Notes: Results are projected to the total population using the survey’s household expansion factor multiplied by the household size. 17 World Bank 2014. 23 When compared to a selected group of European and Central Asian countries the level of social protection benefits relative to household consumption in Croatia is close to the regional average. Croatia’s social protection benefits as a share of consumption of beneficiaries in the poorest quintile, at 43 percent, is 4 percentage points below the group average of 47 percent. However, at almost 42 percent of the total population, this share is almost 6 percentage points above the average of 36 percent. Among the group of countries featured in figure 25, Georgia is the one with the highest benefits level as a percentage of the welfare of the poor (68 percent), but also with one of the lowest benefits in terms of the welfare of beneficiaries in the total population (29 percent). Figure 25. SP Benefits as Percentage of Household Post-transfer Consumption – Selected Countries 70 60 50 43.3 41.8 40 30 20 10 0 Poorest quintile Total Source: World Bank’s Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) Notes: Results are projected to the total population using the survey’s household expansi on factor multiplied by the household size. The poorest quintile is calculated using per capita pre-transfer consumption. Not all high-income countries are included in the ASPIRE database, thus sample is limited to the data available. What is the Impact of SP Programs on Poverty and Inequality? Croatia’s SP programs have a large impact on poverty and inequality. A simulation of how much poverty decreased (poorest quintile) due to the transfers can be done by comparing beneficiaries’ level of consumption recorded in the HBS before and after the social protection transfers. Such simulation estimates that transfers helped reduce the poverty headcount by 57 percent, the poverty gap by 80 percent and inequality (measured by the Gini index) by 30 percent (see figure 26). The size of these impacts depends on several factors, such as the program size (coverage) and the level of benefits. Old-age and disability pensions and allowances are driving large reductions in poverty and inequality, which is not surprising as they are the most generous and contributory in nature. Old-age pensions achieved a reduction of 66 percent reduction in poverty gap, 42 percent in poverty headcount and 18 percent in inequality. However, family benefits also make an important contribution reducing poverty gap 24 by about 26 percent, poverty headcount by about 9 percent and inequality by 4 percent. Small programs such as scholarships, in-kind social welfare benefits have a more modest or even negligible impact. Figure 26. Impacts on Poverty and Inequality reduction due to Social Protection Transfers Social Protection Old-age pensions Disability pensions Family benefits (child allowance, maternity, layette) Family pensions Social welfare in cash Unemployment and retraining benefits Sick leave benefits Social welfare in-kind Scholarships 0 10 20 30 40 50 60 70 80 90 100 Poverty gap reduction Poverty headcount reduction Inequality reduction (Gini index) Source: Croatia's Household Budget Survey 2014. Notes: This figure uses a relative measure of poverty defined as the poorest 20 percent of the welfare distribution (consumption). Percentages of poverty and inequality reduction are calculated as follows: (poverty headcount pre- transfer - poverty headcount post transfer) / (poverty headcount pre-transfer). Same calculations apply for Gini index and poverty gap percentage reductions. The larger the size of the program and the higher the benefit level, the larger the poverty gap reduction is likely to be. Figure 27 illustrates poverty gap reduction (size of the bubble) in relation to the coverage and benefit levels of Croatia’s social protection programs. Old-age pension is the largest program (34 percent of the population), has the largest benefit level on average (48 percent of beneficiary welfare) and has the largest poverty gap reduction (66 percent). Family benefits have relatively large coverage (22 percent) but a lower level of benefits (11 percent of beneficiary welfare), and, thus, poverty gap reduction is lower (25 percent). Family pensions achieve the same 25 percentage of poverty gap reduction with the opposite combination, that is, a higher level of benefit with a lower coverage. Scholarships and social welfare benefits in kind have almost no impact on poverty gap reduction, as their coverage is low (less than 3 percent of the population) and so is their benefit level (5 and 7 percent of beneficiary welfare, respectively). 25 Figure 27. Poverty Gap Reduction due to Coverage and Level of Social Protection Benefits Size of the bubble indicates percentage of poverty gap reduction 60 Old-age pensions Transfers as a share of beneficiary welfare 40 Disability pensions Family pensions (%) Social welfare in cash 20 Sick leave benefits Unemployment and Family benefits Scholarships retraining Social welfare in-kind 0 0 5 10 15 20 25 30 35 40 Coverage of total population(%) Source: Croatia's Household Budget Survey 2008, 2010, 2011, 2014. Notes: This figure uses a relative measure of poverty defined as the poorest 20 percent of the welfare distribution (consumption). Percentages of poverty gap reduction are calculated as follows: (poverty gap pre-transfer- poverty gap post transfer) / (poverty gap pre-transfer). Croatia’s poverty headcount reduction due to social protection transfers is above the average of regional comparators. Social protection transfers have helped reduce poverty by about 50 percent, on average, for a selected group of European countries (see figure 28). Croatia poverty reduction rate is higher at 56.5 percent. Of this group, Romania reports the higher poverty head count reduction (62 percent), while Kosovo reports the lowest (29 percent). Figure 28. Poverty Head Count Reduction due to Social Protection Transfers – Selected Countries 70 60 56.5 50.0 50 40 30 20 10 0 Source: World Bank’s Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE) Notes: This figure uses a relative measure of poverty defined as the poorest 20 percent of the welfare distribution (consumption). Percentages of poverty head count reduction are calculated as follows: (poverty headcount pre- transfer- poverty headcount post transfer) / (poverty headcount pre-transfer). 26 IV. WHAT IS THE PROFILE OF SOCIAL PROTECTION BENEFICIARIES? On average, social protection beneficiaries are almost equally distributed across genders, particularly in the case of social assistance and labor market programs. Figure 29, panel A, shows that about half of social assistance beneficiaries are male, and the other half female. A slightly higher percentage of men benefits from labor market programs (51 percent). On the other hand, a slightly higher percentage of women benefits from social insurance programs (54 percent); this is partly driven by higher life expectancy among women. The age distribution of beneficiaries varies considerably between programs. As expected, about 60 percent of beneficiaries of social insurance are 50 years and older since it includes all the contributory retirement benefits. About 65 percent of beneficiaries of labor market programs are between 15 and 49 years old, or the economically active population. In the case of social assistance, about 65 percent of beneficiaries are between 25 and 49, and between 0 and 14 years old (see figure 29, panel B), a reflection that family benefits and social welfare go to families with children. Only 6 percent of social assistance beneficiaries are above 65, which is at odds with the risk of poverty that this age group faces, as highlighted in the previous sections. The labor market status of beneficiaries also varies based on the type of social protection program. About 59 percent of beneficiaries of social insurance are inactive individuals who are not students (retired). Most beneficiaries of labor market programs are unemployed (38 percent) or employed (31 percent). Interestingly, more beneficiaries of social assistance programs are employed (38 percent) followed by inactive non-students (24 percent). Figure 29. Distribution of Social Protection Beneficiaries by Demographics and Labor Characteristics A. Gender 100 80 48.3 54.1 50.5 60 40 45.9 51.7 49.5 20 0 Social insurance Labor market programs Social assistance Male Female B. Age Groups 100 80 33.7 29.9 60 36.4 25.8 35.6 40 16.0 22.9 20 13.9 28.6 0 Social insurance Labor market programs Social assistance 0-14 15-24 25-49 50-64 65+ 27 C. Labor Market Status 100 23.6 80 58.8 60 37.6 40 20 38.3 24.5 31.3 0 Social insurance Labor market programs Social assistance Employed Unemployed Inactive -student Inactive -non student Source: Croatia's Household Budget Survey 2014. Notes: Figures include direct and indirect beneficiaries 28 CONCLUSIONS Even though on a declining trend, the 2017 at-risk-of poverty or social exclusion rate (AROPE) in Croatia, at over 26 percent, was well above the EU-28 average of 22 percent. One of the most striking facts about the risk of poverty or social exclusion in Croatia is that it affects older people in particular. Among the population above 65, the share of those at risk of poverty and social exclusion in Croatia was in fact almost twice as high as that recorded across the EU-28 (32.7 percent vs. 18.2 percent). As an EU member state, Croatia is a part of the Europe 2020 strategy and, accordingly, aims to reduce the number of persons living at risk of poverty or social exclusion, and inequality within the country. Achieving that goal depends crucially on the effectiveness of the country’s expenditure on social protection programs, and on their design. Croatia spends a significant portion of GDP on social policy. From 2008 to 2017, Croatia spent a yearly average of 71.5 billion Kuna on social protection, or 21 percent of the country’s GDP. Even though expenditure has remained roughly constant in the last years, the 2017 figure (21.4 percent of GDP) was almost two percentage points higher than in 2008, due mostly to an increase in the budgetary allocation to family/children and old age benefits. This share was also lower than the European average (of 28 percent in the EU-28). Two-thirds of the social protection resources are devoted to sickness and health care benefits, and old-age benefits, which are contributory in nature, and another 10 percent is absorbed by contributory unemployment benefits; this, similar to many other European Union countries, leads to a system in which contributory programs dominate. Non-contributory (or social assistance) programs represent a small portion of the overall spending on social protection. Social assistance mostly comprises universal categorical and, to a limited extent, means-tested benefits. Total spending on social assistance reached 3.36 percent of GDP in 2014 - about one-seventh of total social protection expenditure. The largest share of this spending is allocated to veterans’ pensions, which account for almost half of the spending on social assistance programs. These pensions are universal, i.e., everyone in the target groups is entitled to receive them irrespective of their income/means. Child and family benefits account for another 0.80 percent of GDP and are also mostly non-means-tested. Means-tested programs, which include last resort social assistance (i.e. the Guaranteed Minimum Benefit), housing benefits, and in-kind programs, account for less than 1 percent of GDP. In terms of coverage, a large percentage of the total population participate in social protection programs. About 71 percent of Croatia’s population benefits from at least one social protection program. Most importantly, social protection programs almost achieve a universal coverage of individuals in the poorest 20 percent with 96 percent of this population benefiting from at least one intervention. The largest interventions covering the poorest consumption quintile are old age pensions (43 percent), family benefits (41 percent) and disability pensions (21 percent). Cash-based social welfare programs reach only about 15 percent of the poorest quintile. Social assistance benefits are well-targeted; that is, they are, as designed, the most pro-poor among social protection benefits. About 70 percent of social welfare assistance benefits in cash and 65 percent of social welfare benefits in-kind are received by the poorest quintile; this puts Croatia at the top of the performance relative to many other countries in the world. A large percentage of family benefits are also transferred to the poor (41 percent). The portion of in-cash and in-kind social welfare benefits which accrue to the top 40 percent of the population is only 10.4 and 18.5 percent respectively. Contributory pensions are less pro-poor with only 20 to 26 percent of the benefits going to the poorest quintile. This is also to be expected, as social insurance benefits are not targeted based on income but are reserved to social insurance contributors. 29 Croatia’s social protection programs have a large impact on pover ty and inequality. Transfers helped reduce the poverty headcount by 57 percent, the poverty gap by 80 percent and inequality by 30 percent. The programs that reach the most impact are the ones with the largest coverage and the higher benefit level such as old-age pensions, followed by disability pensions. Social welfare benefits, even though well targeted, achieve a smaller overall impact on poverty due to their small size. On average, social protection beneficiaries are almost equally distributed among genders, and the age distribution of beneficiaries varies considerably between programs. About 60 percent of beneficiaries of social insurance are 50 years and older since it includes all the contributory retirement benefits. In the case of social assistance, about 65 percent of beneficiaries are between 25 and 49 and between 0 and 14 years old, a reflection that family benefits and social welfare go to families with children. Only 6 percent of social assistance beneficiaries are above 65, which is at odds with the risk of poverty that this age group faces. In view of these findings, some room to improve exists. First, social assistance programs’ coverage among the poorest 20 percent of the population should be increased. This could be done by paying attention to the demographic groups which face the highest risk of poverty (i.e. the over-65). Second, if politically feasible, the resources devoted to means-tested programs should be increased to achieve even larger impacts on poverty. This does not necessarily have to lead to an increase in the total budgetary allocation to social protection; given the high number of programs in the country, it is in fact likely that some room for consolidation exist to create savings and fiscal space to increase coverage of the poor or most vulnerable. At the same time, this note helped identify some areas which warrant further research. Topics which are somewhat beyond the scope of this work, but could help shed more light on some of the issues raised in the previous sections include: • Programs’ objectives and coverage. As the note argues, there are about 130 national benefits and 3000 benefits managed by cities, municipalities and counties. This points to issue of possible fragmentation of programs and/or duplication. Exploring in more detail the universe of programs, collecting information on individual programs’ objective, coverage, eligibility criteria, and implementation arrangements could provide evidence on whether the overall system is too complex or has some sources of inefficiency. At the same time, because only 52 percent of those in the bottom quintile receive some form of social assistance, delving deeper into the performance of means-tested programs could help establish if and how coverage should be increased. • Drivers of the risk of poverty, especially among women and the elderly. As the note shows, the risk of poverty is higher among women and, ever more markedly, among the elderly. It is possible that these patterns are at least partly explained by difficulties accessing the job market, childcare support, or training opportunities (in the case of women) or by limited access to services or benefits (in the case of the elderly). Narrowing down on categories of people who appear most vulnerable, and better understanding what is driving said vulnerability, could help tailoring policy responses. This is particularly important for the elderly given the rapid ageing which Croatia is undergoing, which will result in a higher number of over-65 in the coming years and will present an even greater challenge down the road. • Lagging Regions. Given the regional disparities in Croatia, further analysis to understand how the analysis presented in this report varies geographically and whether recommendations should vary by region may be called for. 30 REFERENCES ASPIRE (Atlas of Social Protection Indicators of Resilience and Equity). 2019. Database, World Bank, Washington, DC. http://www.worldbank.org/aspire Croatia Bureau of Statistics. 2018. Household Budget Surveys 2008, 2010, 2011, and 2014 rounds. Zagreb. Croatia. Croatia Bureau of Statistics. 2016. Social Protection in the Republic of Croatia (ESSPROS Methodology). Zagreb. Croatia. https://www.dzs.hr/Hrv_Eng/publication/2016/10-01-05_01_2016.htm ESPROSS (European System of Integrated Social Protection Statistics). 2019. Database, Eurostat, https://ec.europa.eu/eurostat/web/social-protection/data/main-tables Handa, S., M. Park. R. Osei-Akoto, B. Davis, and S. Daidone. 2013. “Livelihood Empowerment against Poverty Program: Impact Evaluation.” Carolina Population Center, University of North Carolina at Chapel Hill. Šućur, Z., Babić, Z., Urban, I., & Baran, J. 2016. Struktura naknada, izdaci i korisnici programa socijalne zaštite u Republici Hrvatskoj. Ministarstvo socijalne politike i mladih. Urban, I., M. Pezer. S. Bezeredi. 2017. An Overview of Social Protection Benefits in Croatia. Institute of Public Finance. Zagreb. World Bank. 2014. Social Protection System Modernization Project. Project Appraisal Document. World Bank, Washington DC. 31