FYR Macedonia Special Focus Note: Social Assistance While social protection spending in the former Yugoslav Republic of Macedonia is comparatively high, spending on anti- poverty social assistance programs is low and fragmented across several programs. Furthermore, recent spending has been tilted toward non-means-tested programs, like Parental Allowance, in a context where poverty levels are relatively high and income-generation opportunities for the poor remain scarce, which raises allocative efficiency concerns. The coverage and the adequacy of benefit levels for the Social Financial Assistance, the best targeted program, remains low, while the incentives for social assistance beneficiaries to take up employment remain absent. Gaps in program coverage leave many households mired in poverty and will need to be addressed through reforms. FYR Macedonia Social Protection Spending, while Relatively High, Fails to Address Poverty Effectively FYR Macedonia’s total spending on social protection, including social insurance, social assistance, and labor market programs, is relatively high compared to Western Balkans peer countries, at about 14 percent of GDP; yet anti-poverty programs are among the lowest in the region. While pension spending (old age, disability, and survivor’s pension), at above 10 percent of GDP, makes up most of the total spending on social protection, spending on social assistance—the cash transfers allocated based on the social welfare and child protection laws—amounts to 1.2 percent of GDP. This is among the lowest in the region and below the average level of spending on social assistance among Europe and Central Asia countries which stands at 2.2 percent of GDP. While spending on social assistance has been increasing since 2008, most of the rise was due to categorical rather than well-targeted benefits (Figure 1 and 2). The number of households receiving social financial assistance (SFA) and permanent financial assistance (PFA) — the only explicit anti-poverty programs—more than halved over the past ten years given the low census, and correspondingly, the spending on SFA/PFA dropped from 0.6 percent of GDP in 2005 to 0.2 percent of GDP in 2016. The SFA is means-tested and targets able-bodied persons who cannot meet the subsistence minimum (defined at MKD 2,831 a month in 2016, or below 30 percent of the net minimum wage). This amount increases by a coefficient of 0.37 for each additional household member up to a limit of five. This de-facto guaranteed minimum income (GMI) benefit covers the difference between the minimum subsistence level and the total income of the household. The amount then gets halved after three years of receipt. The PFA targets single unable to work below the poverty threshold. The child allowance (CA) program has also been reduced over the same period, both in terms of coverage and budget (from 0.09 percent of GDP to less than 0.02 percent of GDP). At the same time, categorical benefits increased due to the introduction of Parental Allowance (PA) in 2009. This benefit is provided to mothers who gave birth to a third child. Compared to other social assistance programs, the PA program’s benefit level is very high, at MKD 8,362 per month, and is granted until the third child reaches the age of 10. By 2016, the program’s caseload exceeded 23,500, making up 0.4 percent of GDP. Figure 1: Spending on Social Assistance Programs, 2005-16, Percent of GD Figure 2: Means-Tested versus Categorical Social Assistance Spending, % of GDP Source: Ministry of Labor and Social Welfare FYR Macedonia’s high relative poverty rate of 22 percent in 20161, along with the limited income-generation opportunities available to the poor and vulnerable does not warrant the dramatic shift in spending from means-tested programs toward non-means-tested programs. Despite a decline in poverty in the post-crisis period and improvements in living conditions among the poor, poverty and inequality levels remain high when compared to regional peers and most EU countries (after controlling for economic development levels). 1Also referred to as the at-risk-of-poverty rate, which shows the share of the population living below 60 percent of the median household per adult equivalent income. Anti-Poverty Programs are Well-Targeted, but Coverage Remains Low Social assistance programs cover only one third of the poorest quintile. The result is even more striking considering that the bottom quintile roughly corresponds to the poor population in FYR Macedonia. The social assistance coverage of the lowest quintile increased from 27.2 percent in 2010 to 33.9 percent in 2016, driven by the increased coverage by family and child protection benefits. However, this compares poorly with regional peers which with similar or even much smaller social assistance spending levels have been able to achieve higher coverage than FYR Macedonia—the coverage of the poorest quintile reached levels of almost 70 percent in Bulgaria, 80 percent in Latvia, and 90 percent in Romania. The coverage of the means-tested social assistance (SFA and PFA) of the bottom quintile is even lower. Only 19.3 percent of the people in the poorest quintile receive SFA/PFA which indicates that many low-income households remain uncovered. The SFA/PFA coverage is rather low compared to the coverage of the similar last resort programs in Albania, Armenia, and Kosovo, but it is on a par with other similar programs in the region (Figure 3). Figure 3: Last Resort Social Assistance: Coverage of the Figure 4: Last resort social assistance - percent of Poorest Quintile, %, 2016 or latest year available benefits received by the poorest quintile Source: Europe and Central Asia (ECA) Social Protection Expenditure and Evaluation Database (SPeeD), World Bank Means-tested social assistance is well targeted, as more than 70 percent of all transfers go to the poorest quintile (Figure 4). Furthermore, the benefit incidence for the poorest quintile improved from 57.8 percent in 2012 to 70.3 percent in 2016. The targeting accuracy compares very favorably to those of regional peers. In fact, the leakage of social assistance transfers to the richest quintile (1.4 percent of all benefits) is among the lowest in the region. The accuracy of targeting is also reflected in the fact that leakage incidence to the richest quintile is low in all the social assistance programs except for the disability benefits. Poverty in FYR Macedonia would be much higher in the absence of social transfers. The simulated impact of discontinuing a particular program or combination of social transfer programs shows that in the absence of all social transfers, the poverty rate would more than double, reaching 41 percent. If old-age pensions were discontinued, the poverty rate would grow to 38 percent. On the other hand, the absence of any of the social assistance programs would not impact poverty significantly. The termination of government social and child protection benefits would have only a limited impact (rise in poverty rate to 22.9 percent), but this is due to the small budget allocation to these programs. However, among the social transfers analyzed, the SFA program appears to be the most cost-efficient. For each MKD spent on SFA, the poverty gap decreases by 0.6 MKD. Child and family protection allowances (excluding disability allowances) are next in line in terms of cost-efficiency. One MKD spent on these programs reduces the poverty gap by 0.5 MKD. The other programs, including old-age pensions, show considerably lower cost-benefit ratios. The least efficient transfers are scholarships. Improving Efficiency of Social Assistance in FYR Macedonia The FYR Macedonia government is cognizant of the inconsistencies in the social assistance system and is committed to carry out comprehensive reforms. The key objective of the social assistance system reform should be to increase the effectiveness and efficiency of the system by reforming both targeted and categorical social assistance benefits. Within this process, the government could consider the following reforms within the budget neutral limits: • Discontinuing the Parental Allowance program that is the least efficient while the most generous of all social assistance programs. This could bring savings in the long run and allow resources to be reallocated to programs that target the poor. • Expanding the coverage of the poor while maintaining good targeting performance through the new GMI scheme that would consolidate several existing social assistance programs, introduce improved equivalence scale, and increase the benefit eligibility threshold and the benefit amount. The level of support would equal the difference between the household income and the established threshold. Benefit levels would need to be adjusted regularly to reflect the costs of living. This would offer a comprehensive social safety net for the poor, irrespective of their labor market status or age. An income disregard and a cumulation cap mechanism could be introduced as direct financial incentive to work. The GMI threshold for a four-member family should not be lower than 40 percent of the minimum wage. • The eligibility assessment for the reformed CA should follow the same rules as the GMI program, though the eligibility threshold should be set at a higher level than in the GMI program and should be initiated through a single application procedure. Eligibility should be delinked from the employment requirement. • The current system of disability assessment could be simplified so that potential beneficiaries do not need to go through different sets of medical assessments for each disability benefit, but at the same time a unique disability benefit could be combined with differentiated levels of support based on the associated disadvantages they face. Such disability benefit should not be means-tested.