Journal Issue
Share
Article

Kosovo

Author(s):
International Monetary Fund. Fiscal Affairs Dept.
Published Date:
May 2016
Share
  • ShareShare
Show Summary Details

I. Introduction

1. Kosovo is unique in many respects, it is the demographically youngest country in Europe—and is also one of the poorest in the continent. Both unemployment and inactivity rates are high—at close to 30 and 60 percent, respectively, with youth unemployment at particularly worrisome levels. Historically, Kosovo was a major source of migrant workers but the conflict in the late 1990s as well as economic difficulties before and after the conflict resulted in large-scale outflows of working age people. Consequently, remittances and regular visits by members of the diaspora play a crucial role in supporting families and providing seasonal boosts to domestic consumptions. Remittances also make it more difficult to effectively identify families in need and have a distortive effect on the structure of fiscal revenues.

2. Poverty is pervasive, with close to one-third of the population living on less than the national poverty line and over 10 percent surviving on less than the extreme food poverty equivalent. Poverty is the main policy challenge of the country which can be addressed through employment-friendly economic growth and pro-poor social spending. While economic growth has been impressive—averaging at 3.6 percent between 2008 and 2015, Kosovo’s track record remains poor on both job creation and poverty alleviation. Unemployment, inactivity and poverty rates have been stagnating for almost a decade. Population growth and a young age structure will maintain supply-side pressure on the labor market which makes it imperative to provide strong incentives to employment. The need to protect the poor while promoting employment in a fiscally constrained environment forces the government to make difficult policy choices.

3. The current system of social cash transfer represents a departure from the fiscally sustainable and conceptually homogenous structure of the early 2000’s and is highly inequitable. The years since independence in 2008 were characterized by a proliferation of budget-financed cash transfer programs which allocate a growing portion of the social budget on the basis of political bargaining power and less on the basis of need. War-related schemes tend to pay significantly higher benefits on the basis of significantly less stringent criteria than programs explicitly designed to assist the destitute (such as social assistance or the standard disability pensions). The coverage of social assistance is low and benefit levels are inadequate—while working age people with limited or no health-related capacity limitations draw benefits which are designed in a way discouraging employment. Over the past years, there has been increasing pressure to re-introduce earnings-related pension schemes. Some of these benefits are based on contributions made prior to 1999 and do not assume new entitlements (such as the contributor pension scheme), while others are designed as permanent schemes. The common feature of these earnings-related pension schemes is that they are noncontributory, recognizing pre-independence or new entitlements, without assets or revenues to offset the liabilities thus generated.

4. The system of social protection cash benefits needs to undergo substantial reforms to improve its equity and sustainability. The objectives of this report is to provide reform options to reduce the level and growth rate of special benefits and, partly through the fiscal space thus obtained, to improve the coverage and adequacy of social assistance. The rest of the report is structured as follows: Section II discusses age and disability pensions, Section III covers special benefits, and Section IV reviews the social assistance schemes. All three sections provide an overview and presentation of issues and, finally, a list of recommendations. Section V briefly presents policy and administrative issues which impact on all three major social expenditure categories. Finally, five technical appendices provide regulatory comparisons across existing old age and disability schemes, offer comments on a policy document and a draft law shared with the mission, and quantify the expected fiscal impact of selected measures.

II. Age and Disability Pension

A. Overview

5. The pension landscape in Kosovo has rapidly expanded and at present includes numerous programs, covering both elderly and working age people. Schemes range from funded defined contribution pensions (KPST) to universal, flat, noncontributory provisions (see Appendix I). This section provides an overview of the pension schemes governed by the Law on Pensions Financed by the State” (hereafter “2014 law”) but excludes other benefits also termed “pensions” payable to special groups which are described in the next section.1

6. Age and disability pensions are the largest social cash benefit programs with outlays of 3.6 percent of GDP—more than two thirds of all social protection spending. Figure 1 provides the breakdown by main benefit types: basic pension, contributory pension and disability pension. The observed rather sharp increases in basic and contributory pension expenditures from 2011 to 2014 have been driven by two factors. First, steep increases in the real value of monthly benefit amounts—with relative increases being strongest for basic pensions (as summarized on Figure 1). Second, a substantial increase in the number of basic and contributory pension beneficiaries of approximately 25 percent during the period 2011–2015, while numbers of disabled have stagnated over the same time-span.

Figure 1.Pension Expenditures, 2010–2015

(Percent of GDP)

Source: Ministry of Labor and Social Welfare

7. The various pension benefits organized by the 2014 law are mutually exclusive. The 2014 law also extends the one-beneficiary—one benefit rule to other cash transfers administered by the Ministry of Labor and Social Welfare (MLSW) but remains silent on schemes managed by other government bodies. Some controversy is also introduced by regulations defining mutual exclusivity across pensions financed by the state, while others refer to pensions administered by MLSW, although in the first years of independence, the two notions were equivalent as no schemes were managed outside MLSW, this is no longer the case. With the gradual emergence of budget-financed occupational pension and cash benefit schemes, the effectiveness of the exclusivity rules is eroding. Examples are the schemes for the Kosovo Protection Corps (KPC), the Kosovo Security Force (KSF) and the Police.

8. The basic pension is a general-revenue financed universal flat benefit, covering all citizens aged 65 and over without other pension income. The scheme is administered by MLSW, providing benefits to 132,000 people at EUR 75 per month with a current cost of about 2.1 percent of GDP per year. In addition to age, pension receipt is also conditional on residence and drawing no other budget-financed pension administered by MLSW. To ensure continued benefit receipt, individuals have to present themselves every 6 months in person in an office designated by the ministry.2

9. The contributory pension is a budget-financed defined benefit scheme for citizens above age 65 with a sufficiently long work experience prior to 1999. Eligibility requires at least 15 years of contributions prior to 1999 into the social security scheme of Yugoslavia.3 Basic and contributory pension cannot be paid to the same person. Eligible individuals were, until 2015, entitled to a flat benefit of EUR 140, with the same residence and administrative procedures in place as for the basic pension.

10. In establishing eligibility for contributory pensions, the 2014 law recognizes various noncontributory periods. The qualifying periods taken into account for the career requirement now also include years of work in the parallel (informal) health, education or “other” sectors in Kosovo between 1989 and 1999.4 In addition, the law grants all individuals who can prove 15 years of qualifying affiliation—thus currently all beneficiaries of the contributory scheme—an extra 25 years of notional earnings history, rewarded with a 0.5 percent benefit increase per service year added5. In terms of benefit levels, this provision equates people who actually contributed for 40 years prior to 1990 to those who only did so for 5 years but achieved the same 40 years of service by the law’s recognition of additional 35 years without contributions. As a result of the additional 25 years’ recognition, benefits increased by 12.5 percent in January 2016, from EUR 140 to EUR 158. While no cost estimates are yet available for the first measure because of a lack of data on the scope of additional qualifying periods of employment in parallel (informal) health, education and “other” systems, the second measure will generate additional expenditures of approximately EUR 8.7 million in 2016 (0.2 percent of GDP).

11. The 2014 law also introduced benefit differentiation according to pensioners’ level of education, using educational attainment as a proxy for past earnings. While policymakers’ admitted objective was to restore benefits’ earnings-related nature, an explicit link with prior earnings was not possible given that not all records are available. According to MLSW, four levels of education will be differentiated, with benefits paid to people with education not exceeding primary schooling to remain unchanged. Although the multipliers to be applied to benefits paid to pensioners with higher levels of education are not yet determined, the recurring fiscal costs may be substantial. Table 1 provides a static simulation of the potential budgetary impact based on the observed numbers of beneficiaries in December 2015, estimating the ensuing fiscal cost at EUR 7.4 million (0.1 percent of GDP) based on a set of hypothetical multipliers Note that the fiscal cost in future years also depends on the scope of the eligibility expansion resulting from noncontributory service time and the possible changes to the education mix in the pensioner population.

Table 1.Illustrative Full Year Fiscal Cost of Education Link of Contributory Pension7
Education levelShare (Percent of beneficiaries)Number of beneficiariesMultiplicative adjustment factorMonthly benefit (EUR)Annual incremental cost (EUR)
No formal education5020,18211580
Secondary3514,1281.11742,712,576
Post-secondary72,8261.32051,593,864
College/University83,2291.52373,061,092
Total10040,3657,367,532
Source: MLSW, and IMF staff estimates.
Source: MLSW, and IMF staff estimates.

12. Disability pensions are available for fully disabled resident citizens aged between 18 and 65. The scheme is financed from the budget and is operated by the MLSW, with no benefits available for partially disabled individuals—i.e., individuals with a disability level below 100 percent. The uniform monthly benefit is EUR 75, equal to the basic pension. The severity of the disability is ultimately determined by the Medical Commission of the MLSW, with reference to the notion of ability to work, with a recurrent re-assessment after 1, 3 or 5 years—with the exact periodicity for each individual case being decided by the Medical Commission. The age cutoffs for the scheme are justified by the fact that disabled children under the age of 18 are covered by a separate benefit, whereas at age 65 the basic pension becomes accessible.6

13. The 2014 law introduced a separate system of work disability pensions. The scheme is financed by the state and is not yet implemented in practice. It covers those individuals younger than 65 who received a similar benefit prior to 1999, as well as employees whose onset of a work-related disability occurs after this date.8 Most details of implementation—including qualifying disability levels—are relegated to administrative regulations which are not yet available. The monthly benefit amount is set equal to the basic pension (EUR 75). For pre-1999 beneficiaries the eligibility conditions of the time remain applicable, whereas for post-1999 work accidents or professional diseases only fully disabled qualify for benefits.9 Contrary to single decision-making body formed by the Medical Commission of the MLSW for the permanent disability scheme, there is a second decisive actor for work disability pensions: specialized public institutions of labor medicine operating under the authority of the Ministry of Health will verify and evaluate the insurable events even before the application files are transmitted to the Medical Commission of the MLSW.

14. According to the 2014 law family pension benefits are available to surviving spouses and dependent children based on social insurance entitlements accrued prior to 1999. These benefits, similarly to the work disability pensions described in the previous paragraph, are not yet paid. In principle, the system will grant family benefits in three situations: first, to individuals who were in receipt of family pensions before 1999; second, to those who become eligible as a result of the death of a worker who had been insured under and contributed to the system prior to 1999 for at least 15 years; and third, to those who become eligible for benefits as a result of a death caused by work accident or occupational disease irrespectively of the effective duration of affiliation before 1999. The benefit for surviving spouses is the same as the basic pension, with a 20 percent increase for each dependent child (below 18, or 26 in case of continued studies). Benefit receipt while in employment—including self-employment—is prohibited.

15. In addition to the above schemes and benefits, former workers of the Trepca mining complex can receive an early retirement pension under certain conditions. The 2014 law formalized and regulated previous ad hoc regulations and practices. The aim of the program is to provide workers who were involuntarily laid off in the 1990’s with an early retirement option. Eligible applicants must be resident citizens, at least 50 but no more than 65 years of age, have a recognized work history of at least 10 years in the mining sector, absence of any form of employment or self-employment income and suffer from a work-related disability of at least 50 percent.10 Qualifying experience for the Trepca early retirement pension includes periods of actual employment as well as periods of unemployment following involuntary layoff. The monthly benefit amount is currently set—by way of administrative regulation—at 105 EUR, with individuals rolled over into the regular age pension schemes upon reaching age 65.11

16. All the above pension benefits should be adjusted annually for prices—however, the law allows flexibility to the government, in reflection of available fiscal space. The cost of living index is established by the Ministry of Finance. The Ministry of Finance also computes the necessary adjustments in the pension amounts, and then proposes changes to the Government that ultimately decides on the magnitude of the granted increase. In practice, no such “automatic” indexation has taken place since the system’s inception. Rather, ad hoc adjustments have occurred—with only a loose connection to current or cumulative inflation. Figure 2 provides a comparative view on benefit amounts and consumer inflation over a 10-year period.

Figure 2.Benefit Levels and CPI, 2006–2015

(Year-over-Year Percentage Change)

Source: MLSW, and MoF.

17. The public pension schemes are complemented by mandatory pension savings in a publicly managed, defined contribution scheme, the Kosovo Pensions Savings Trust (KPST). Contributions are levied on all employment and self-employment income at a (tax-deductible) rate of minimum 5 percent each for employees and employers.12 Accumulation of assets takes place in individual accounts. Regular old-age benefit payouts start at the age of 65, with early withdrawals possible in case of 100 percent disability that is recognized under the public pension scheme. For old-age benefits, the law provides for two payout mechanisms: annuitization and phased withdrawals—the latter serving as a fallback mechanism if no financial actor is willing or able to provide such annuity products. Given small retirement account balances, which are not expected to increase significantly in the coming years, no meaningful life annuities are possible and all payouts take the form of phased withdrawals. Since average account balances at retirement are approximately EUR 4000 and withdrawals cannot be less than EUR 150 per month, KPST balances can provide an income stream for a period of 2.5 to 3.5 years on average, depending on net returns realized during the payout phase.

B. Issues

18. The number of beneficiaries older than 65 exceeds official population figures by at least 20 percent in every age bracket, requiring urgent attention. Figure 3 illustrates the evolution of the basic and contributory beneficiary population; it also documents that the number of beneficiaries aged 65 and older outstrips the resident population in the same cohort, with an increasing trend particularly driven by the number of base pension beneficiaries. Conceptually, four main reasons could explain this discrepancy. First, there could be individuals claiming both basic and contributory benefits: according to the authorities, such double claiming is technically not possible given administrative procedures in place. Second, population statistics from the census could be very unreliable: there has so far only been one census, furthermore with incomplete territorial coverage. Third, information on deaths could be imperfectly recorded as a result of nondeclaration of deaths to the competent authorities—however, this can only explain the level and not the change in the extent of the discrepancies. Fourth, and foremost, the concept of residency might be deeply flawed, both for benefit eligibility and for the census. Many beneficiaries qualifying as residents might actually not be habitual residents as captured by the census, while the census itself likely struggles with issues around the notion of residency linked to civil and citizenship matters.

Figure 3.Beneficiaries of Basic and Contributory Pensions vs. Population 65+

Source: MLSW, and the 2011 Population and Housing Census.

19. There is a strong gender dimension to the observed benefit claiming patterns. Figure 4 illustrates the age distribution of pensioners and the population 65 and older for the year 2015. Three features stand out. First, approximately 95 percent of female pension beneficiaries above 65 claim the basic pension. The respective proportion for men is 56 percent—while the remaining 44 percent claim the higher, contributory pension. The difference is likely driven by shorter qualifying careers. Second, the observed wedge between the beneficiary population and the overall population older than 65 in Kosovo is stronger for men than for women, underlying the role of nonresidency. Historically, men were more mobile while the two other factors behind the discrepancy between population and beneficiary indicators do not systematically differ across genders. Third, while for men the share of contributory pensioners is approximately constant across age groups, there is a noticeable increase in the share of female contributory pension recipients among younger cohorts. The above situation has two important policy consequences. From a fiscal perspective, this trend may provide an additional driver for the growth of pension expenditures beyond sheer cohort size. From a welfare perspective, any reform of the basic pension will have a larger impact on women than on men.13

Figure 4.Pension Beneficiaries vs. Population by Cohort and Gender

Source: MLSW, and the 2011 Population and Housing Census.

20. Trepca and family pensions paid to individuals below age of 65 have potentially severe labor market implications. Conceptually, they resemble selective and permanent unemployment benefits, discouraging employment and even exiting the labor market in order to claim benefits. Allowing beneficiaries to generate labor income—with a benefit claw-back mechanism should help address these concerns at least partially. Further active labor market policies should be implemented to encourage these individuals to re-enter the labor market.

21. The current design of family pensions creates strong inequities and inefficiencies. The fact that family benefits are only selectively granted to some categories of beneficiaries such as families of war-related pensioners, contributors to the old (former Yugoslav) system but not of “regular” current contributor is hard to rationalize, particularly in light of the general budget financing. 14 The adverse distributional consequences are further reinforced by the choice of essentially granting lifelong benefits to spouses instead of offering transitional assistance, in line with international practice, designed to help adjusting to the new financial realities but without discouraging employment. As for Trepca pensions, the categorical earnings test likely excludes many eligible surviving spouses from the labor market.

22. The conditions of the permanent disability and work-related disability pensions are unreasonably strict, leaving partially disabled individuals uncovered. Only individuals qualified as 100 percent disable can qualify for benefits. The situation contrasts with war-related or special categories that can benefit from partial disability schemes allowing them to combine benefits with partial labor market attachment. An expansion of the system to include partial disability coverage could be warranted, with appropriate care taken to create minimal work disincentives and ensuring continued labor force attachment.

23. The lack of a consistent and labor-market oriented approach for evaluating disability across programs is problematic. Though all relevant disability thresholds are expressed in terms of percentages, there is no common set of medical criteria used—neither is the remaining ability to work consistently taken into account Finally, the apparently unsystematic dispersion of evaluation and decision powers between general practitioners, specialists, and ministerial medical commissions clearly leads to inefficiencies (and likely inequities) in procedures and likely also in outcomes.

C. Recommendations

24. Restrict payment of noncontributory age and survivor benefits to resident citizens.

  • Require all recipients of noncontributory age and survivor pensions between the ages of 18 and 65 to present themselves at designated government agencies once every two months;

  • Require recipients of age and disability pensions over age 65 to present themselves once every three months;

  • Limit retroactive collection of benefits to one month; and

  • In case of fully disabled or hospitalized recipients, introduce visits by CWS officers or other government agencies.

25. Limit multiple benefit claims.

  • Clarify ambiguous regulations and make all pension benefits financed by general government mutually exclusive.

26. Establish uniform disability definitions, scales, procedures and eligibility criteria.

  • Legislate a uniformly applicable list of medical conditions and the rules determining the manner in which multiple conditions are compounded;

  • Re-define eligibility for disability-related benefits on remaining capacity to perform any type of work;

  • Legislate a new set of eligibility levels: make the notion of partial disability universally applicable with eligibility for cash benefits established at no lower than 70 percent; and

  • Introduce differentiated disability pension benefits, with benefit levels conditional on disability levels, ranging, for instance, between 0.7 and 1 times the basic pension—possibly complemented by ad hoc social assistance benefits.

27. Initiate the re-certification of all recipients of benefits conditional on disability.

  • Re-certify all recipients according to the new, uniform criteria, including the application of the common criteria for a partial disability;

  • Establish procedures for notification, re-certification and appeals; and

  • Contract independent medical experts to be employed in the process of ad-hoc supervision of medical commissions and appeals.

28. Time-limit and harmonize family benefits.

  • In case of working age survivors, limit eligibility to noncontributory family benefits to two years, regardless of the nature of the deceased main beneficiary’s noncontributory pension; and

  • Harmonize benefit levels at EUR 75 per month, with a grandfathering period of one year for survivors currently receiving a higher benefit.

III. Special Benefits

A. Overview

29. There is a broad range of cash benefits available to special groups, typically based on participation in the independence struggle, suffering from the consequences of political persecution, or the war. Currently, the majority of these benefits is paid to people younger than the retirement age—however, most of these benefits continue to be paid after recipients reach the normal retirement age of 65 years. The regulations governing these benefits also provide for survivor benefits to spouses, children, and parents.

30. Social protection expenditures are heavily skewed towards benefits unrelated to poverty. In 2015, expenditures on merit-based (war-related) and occupational benefits represent 15 percent of total social protection expenditures and approximately 40 percent of outlays benefitting working age recipients. From 2016 on, the system’s inequity are expected to deteriorate further: with the roll-out of the new War Veterans’ Pension, merit-based benefits will grow to approximately 55 percent as a share of payments to working age people.15 In terms of expenditures, the five largest items were benefits payable to Kosovo Liberation Army (KLA) war invalids, martyrs’ families, the families of civilian victims of the war, former workers of the Trepca complex and war veterans. From 2016 on, the single largest item will be war veterans’ pension, expected to cost on an annualized basis approximately EUR 36 million in its first year of operation—as much as the combined expenditures on the four other large expenditure programs and well over the outlays on the means-tested social program.

31. Most benefits do not preclude employment and generate derivative (survivor) benefits which are not conditioned on either means or unemployment. There are differences, too, however: most importantly, the new war veterans’ pension will require that recipients be registered as unemployed. Payment of merit-based benefits continues after retirement and, in light of beneficiaries’ age distribution, will represent a fiscal burden for at least two more decades.

32. The most recent and largest war-related benefit is war veterans’ pension which is expected to cover approximately 18,000 people with an additional 12,000 who may, potentially, become beneficiaries at a later time. In order to be eligible, applicants’ former KLA affiliations must be certified by a special commission; they cannot be employed or self-employed and must be younger than the normal retirement age of 65. Currently, 13,000 cases have been completed with benefits awarded and another 5,000 applicants are awaiting the final settlement of their cases, with most of them expected to be successful. A further 12,000 people have been issued with documents proving their former affiliation to KLA who may, at a later a date, meet the remaining eligibility conditions and start drawing benefit. The benefit level is explicitly linked to the statutory minimum wage, currently at EUR 170. Once the recipient becomes eligible for the basic (or contributory) pension, the total amount continues to be paid if needed in the form of a gap-filling “pension supplement” to other pension benefits. War veterans’ survivors are eligible for a family pension at 70 percent of the deceased person’s pension, payable until the family member’s retirement (in case of spouses), 18th or 26th birthday (in case of children and depending on how long they attend school full time).

33. Survivors of people who died or disappeared during the war can draw family benefits, the amount of which depends on whether the deceased person was a KLA member or a civilian and on the number of family members lost in the conflict. The families of martyrs (KLA fighters who died during the armed struggle or later, as a direct result of injuries sustained) receive benefits of between EUR 448 and 672 per month (approximately 6-to-9 times the basic pension), depending on the number of martyrs. Currently, less than 2,000 families receive this benefit with expenditures expected to gradually decline. The families of civilian victims of the war receive lower benefits (between EUR 169 and 253 in 2015) but the number of recipients is higher, at 4,700 families. Both group of beneficiaries are permitted to draw other nonpension benefits in addition to their war-related family benefits and are also allowed to have labor income. Following retirement, benefits continue in addition to the basic and contributory pensions. Families of missing KLA and civilian persons receive similar treatment as those of martyrs and civilian victims, respectively.

34. People who suffer from disabilities attributable to the war are eligible for generous KLA war invalids’ or civilian war invalids’ pension; surviving family members receive family pensions—with less strict eligibility conditions and considerably higher benefit levels than in the case of regular disability pensions. Eligibility for war invalids’ pensions is conditioned on a level of disability in excess of 20 percent. This criterion compares very favorably to the full (100 percent) disability condition required for social security disability benefit. Furthermore, the lowest KLA war invalids’ pension—awarded to people with 20 percent disability—is three times as high as the normal disability pension, at EUR 225 per month.

35. The distribution of KLA war invalids’ pension reveals that the overwhelming majority of beneficiaries is in the lowest disability categories. These levels of disability suggest medical conditions which were difficult to verify objectively at the time of claim and which should not, per definitionem, exclude employability. While the number of KLA war invalids is small—less than 4,000—this expenditure item was the largest, until the recent introduction war veterans’ pension, among benefits payable to working age recipients. Civilian war invalids’ claims are contingent on a disability of at least 40 percent and their benefits are lower than those of KLA war invalids—the lowest benefit is still more than twice as high as that received by regular disability recipients suffering from full disability.

36. In addition the above benefits, a new scheme for victims of sexual abuse was also legislated in 2014. The scheme is not operational yet, as the eligibility commission has not been established and benefit levels have not yet been defined by the government.

37. The pension schemes benefitting the members of the (now defunct) Kosovo Protection Corps (KPC) and its successor, Kosovo Security Force (KSF) differ markedly from other schemes and bear a strong resemblance to civil service schemes in non-European emerging economies. Both schemes operate as noncontributory, earnings-related final salary arrangements, although some differences exist: most importantly, whereas the KPC pension scheme will, eventually, wind down as there are no new entrants into the scheme, the KSF scheme will remain open.16

38. Eligibility for a KPC pension is relaxed and benefits are generous. Participation is conditional on having served at least for 5 years in KPC (during its 10 years of existence) and reaching age 45. A full KPC pensions are available from age 50 at 70 percent of the retirees’ last salary received from KPC. Former members can also retire early, at age 45 in which case there benefit will be set at 50 percent of their final salary. KPC pensions are payable past age 65, in conjunction with the basic and contributory pensions. KPC pensioners’ survivors are eligible for a family pension set at 60 percent for spouses and 20 percent for children (irrespective of the number of young survivors). There are approximately 900 KPC pension recipients, with 55 percent younger than 65 and 10 percent paid a family benefit.

39. KSF pensions are a small but generous program financed by the budget, paying final salary pensions. Full pensions are payable from age 50 or 55 (officers) with at least 20 years of service. Benefits are set at 60 percent of final salaries: although the scheme has a benefit formula (40 percent plus 2 percent per year of service), the service time criteria ensures that pension will be no lower than 60 percent while an explicit cap limits pensions at no higher than 60 percent. Former KPC members who continued service in KSF can retire at any age after 3 years of service in KSF. In this case, if retirement happens below age 50 or 55 (in case of officers), then pension will be paid at 40 percent of the final salary, increasing to 60 percent at reaching 50 or 55.

40. In addition to the cash transfers, various other quasi-cash, in-kind and administrative benefits are available to beneficiaries of the above schemes. These benefits include utility subsidies, reduced health care and education co-payments, free or subsidized access to medical appliances, duty and tax easements when importing vehicles, preferential treatment of family members’ university applications, etc. With the exception of electricity subsidies, the cost (including tax expenditures) of these programs are not reported separately and are not considered in establishing eligibility for cash transfers.

B. Issues

41. The current system of numerous war-related benefits is fiscally costly and inequitable. In 2015, the share of recipients of war-related benefits was 6.6 percent of the total cash transfer recipient population, while expenditures on war-related benefits was disproportionately higher, at 18 percent (Figure 5a and 5b). Eligibility criteria are unrelated to economic need, effort to find employment and, by covering family members, expands considerably both the timeframe of these schemes and the public expenditures required to finance them. Given the poor coverage and inadequate benefit levels of the social assistance scheme, it is crucial to reform these schemes and apply the fiscal space thus created towards the people living in extreme poverty.

Figure 5a.Beneficiary Composition

(Percent of total individuals receiving social protection benefits)

Source: MLSW

Figure 5b.Expenditure Composition

(Percent of total social protection spending)

Source: MLSW

42. The newly legislated war veterans’ pension is expected to become the most expensive special benefit program which will further deepen the safety net’s inequities. The total cost of the program is expected to reach approximately EUR 35 million in 2016—40 percent more than originally predicted—but further expenditures may be generated by people who meet the basic eligibility condition of verified KLA affiliation and who may become unemployed, possibly through their own volition. The theoretically possible upper limit of expenditures on this program is in excess of EUR 60 million (1.2 percent of GDP) and while some substitution across the new and existing benefits may lower both the theoretical maximum and the realized expenditures, this effect is unknown and should not be the basis of conservative expenditure policies. Thus, the schemes will need to be substantially redesigned if its fiscal costs are to be curtailed.

43. War and civilian war invalidity benefits are based on substantially more permissive disability criteria and provide generous benefits in comparison to standard disability benefits. The amount war invalidity benefits, based on invalidity levels between 20 and 100 percent, ranges between 3 and 6 times the benefit available to fully disabled people outside these schemes. Civilian war invalids’ eligibility starts at 40 percent disability levels and benefits range between 1.6 and 2 times standard disability benefits. Furthermore, and perversely, while fully disable persons receiving EUR 75 per month are requested to regularly submit themselves to re-certification, war invalidity pensioner with significantly lower levels of invalidity only had to be assessed once, at the time of application. Last, the conditions under which war invalids were assessed are unknown: there are no assurances that current beneficiaries would meet eligibility conditions if they were measured along the same scales as standard disability pensioners.

44. Family pensions are currently only available to survivors of war-related beneficiaries, at levels of generosity well beyond those that will prevail under the new, family pension scheme. Going forward, and in contradiction with international practice, the risk of losing a family’s breadwinner is still not covered for the wider Kosovar population. At the same time, the survivors or martyrs, civilian and former KLA war invalids, KLA members and civilians who disappeared in the war and war veterans are entitled to family benefits. Eligibility is life-long for spouses and up to age 26 for children. The benefits can range between EUR 96 per month (civilian war invalids’ survivors) to EUR 672 (missing KLA members’ and martyrs’ survivors). Similarly to other permanent and relatively generous benefits, war-related family pensions discourage employment and introduce inequities into the social safety net.

C. Recommendations

45. Reduce expenditures of war veterans’ benefits, while ensuring greater individual choice.

  • Limit total benefit receipt at EUR 4,000, payable in instalments not exceeding EUR 120 per month, over a period and frequency of the beneficiary’s choice; and

  • Permit applying remaining eligibility balance as collateral for establishing micro and small enterprises.

46. Discontinue the direct budget financing of earnings-related pension schemes and introduce parametric adjustments improving their equity and sustainability.

  • Conduct an actuarial assessment to establish the long-term equilibrium contribution rates required for financing earnings-related benefits accruing under the KSF pension scheme;

  • Make KSF pension contributory, equally sharing the contribution burden between employees and employers;

  • Revise the KSF pension scheme’s benefit formula and expand the pension calculation base from the last salary to a full career average;

  • Replace rank-indexation of benefits with price indexation; and

  • Observe the same provisions when establishing other earnings-related schemes, including the planned substantial expansion of the police pension scheme.

IV. Social Assistance

A. Overview

47. The social assistance scheme (“Asistenca Sociale”) is the only cash social protection program in Kosovo that was designed with the specific objective of poverty reduction among the working age population. It is a budget-financed guaranteed minimum income program, which was implemented in 2000. Around 26,000 families (or 107,000 individuals) benefited from the social assistance scheme in 2015. The average size of the beneficiary families was around four people per family, and the average benefit amount was about EUR 81.2 per month per family.

48. Social assistance is the second largest program in terms of the number of beneficiaries, but its share in the total social protection spending is small. Among the social protection programs funded by the Ministry of Labor and Social Welfare’s (MLSW) budget, social assistance beneficiaries accounted for around 43 percent of the total number of beneficiaries, and its beneficiary share remains rather constant over the last decade. On the other hand, their budget share is small and has declined over time (Figure 6). By end-2015, spending on social assistance was about 8 percent of total social protection spending.

Figure 6.Number of Social Assistance Beneficiary Households and Spending, 2006–2015

Source: MLSW.

49. The social assistance scheme is administered by municipal Centers for Social Work (CSW). Upon the decentralization of social services and social assistance in 2009, the CSWs became the responsibility of municipalities and the respective municipal division is the Municipal Department of Health and Social Welfare. Two main sub-divisions of the CSWs are responsible for the social assistance scheme and social services. On social assistance, the CSWs continue to handle applications and the process of verification and validation. However, financing, payment decisions, and actual transfers remain centralized, primarily under the MLSW.

50. An application for social assistance is valid for at least six months and not more than one year—provided that the family continues to satisfy the eligibility criteria during this period.17 The application and continuing control procedures are as follows. At the beginning of the application process, the applicants need to submit a set of documents to the CSWs to determine eligibility, including proofs of identity and marital status for all household members, unemployment certificate, certificate of disability, land cadaster and other proofs of asset ownership, school certificates, tax registration and insurance documents, and other documents deemed relevant by the CSWs. In the next stage, CSW staff conduct interviews and home or neighborhood visits. The list of eligible families is subsequently forwarded to the MLSW, which makes the final decision on eligibility and transfers.

51. The residency criterion is strictly enforced in case of social assistance recipients: as part of the continuing control over eligibility, social assistance payments are made on a monthly basis by the MLSW. If a transfer is not collected from the special bank account—opened by MLSW for the beneficiary—within two months, the benefit is withdrawn. Families are also encouraged to declare to the CSWs if they receive income or transfers in any month, so that the transfers of social assistance benefits can be halted temporarily. They could request for an exemption if these income and transfers are used for some exceptional purposes such as medical necessity. The MLSW also has a right to conduct ad-hoc inspections and can cross-check reported income and transfers using information from tax administration offices and the KPST. If the missed report income triggers a disqualification for benefits, the family will be ineligible for social assistance until the next application round (six months to one year—depending on the social assistance category).

52. The program is targeted using categorical eligibility criteria, as well as means testing. Two categories of families are eligible for social assistance: Category 1 beneficiaries are families with all members being dependent and hence not capable of working.18 Families applying on the basis of one or more family members being permanently disabled and thus dependent will be required to have such member’s medical conditions reviewed by the Medical Commission. Category 2 beneficiaries are families with only one member capable of working, who is registered as unemployed at the Employment Office and is actively seeking for work (through the Employment Office), and with at least one child under the age of five or an orphan under the age of 15. An additional family member registered as unemployed will cause a family to be ineligible. In addition to these categorical criteria, families must also pass an asset test—such as no ownership of motor vehicles or land larger than 0.5 hectares.19

53. The benefit level of social assistance has improved but is still low. The monthly amount of social assistance paid to eligible families is determined by the difference between the family’s assessed income and the Monthly Gross Standard Rate (MGSR). The MGSR varies with family size, ranging from EUR 50 to EUR 150 per month. In addition, beneficiary families receive a child allowance of EUR 5 per month for each child under the age of 18. The MGSR has been adjusted only three times, in 2003, 2009, and 2015, since the program’s implementation in 2000. Due to legislative amendments in 2012 with the objective of improving the benefit level, and the 25 percent increase of the MGSR in 2015, the average monthly amount of social assistance increased from EUR 66.1 in 2010 to EUR 81.2 per family in 2015. For a family with four members, which is the average size of beneficiary families, this implies a per person increase of 13 cents per day over the five-year period—from EUR 0.54 in 2010 to EUR 0.67 per family member per day in 2015.

54. While the benefit level slightly increased, the number of beneficiaries and total social assistance spending declined due to stricter eligibility criteria and tighter enforcement in recent years. Category 2 eligibility has been tightened since early 2013 by including only families with a sole unemployed member who registers oneself as being unemployed and looking for work at the Employment Office. The sole able-bodied member is required to obtain the unemployment certification from the Employment Office, which is valid for six months. At the same time, the implementation and enforcement of all existing eligibility criteria have become stricter—for example on screening, cross-referencing information, and inspection. There have also been improvements in monitoring sources of income and asset ownership. As a result, between 2010 and 2015, the number of beneficiary families decreased by around 10,000 households, and spending on the social assistance scheme declined from 11.3 to 8.3 percent of total social protection spending (Figure 3). As a share of total social protection spending, spending on the social assistance scheme fell significantly from 17 percent in 2010 to 8 percent in 2015.

B. Issues

55. While the targeting efficiency of the social assistance scheme is good, its coverage is poor, leaving many poor households unassisted. In 2011, 54 percent of total spending on social assistance went to the population in the bottom quintiles, compared with Eastern Europe and Central Asia’s average of 36 percent (and the Balkans’ average of 39 percent). There is a relatively low leakage to the top quintile—only around 3 percent of social assistance spending. However, less than one-third of the poor population (living with less than the national poverty line of EUR 1.72 a day) received the benefits—10 percentage points lower than the regional average. Furthermore, the average per capita benefit received by the bottom quintile was only around EUR 0.55 per day.

56. The social assistance scheme relies on inefficient proxies of poverty as eligibility filters. The current eligibility criteria exclude many poor families, compromising the scheme’s poverty alleviation objectives (Figure 7). Furthermore, some of the filters—such as the lack of financial and productive assets—are exceedingly difficult to effectively verify. In addition, social assistance is less accessible to those who are able to work and this could have a significant impact on labor market behavior, which in turn has an adverse impact on poverty.

  • The age limit of the youngest child. It is unclear why Category 2 beneficiaries should become ineligible when the age of their youngest child reaches five. While one could use, for instance, compulsory school starting age to justify an age limit, the cost of raising a child does not become significant less and/or discontinuous at the age of five. In addition, children are more likely than others to be in poverty as their overall poverty rate in 2011 was 32.8 percent compared to 29.7 percent for the entire population (World Bank, 2013).20 Families with many children are also associated with a greater likelihood of being in poverty and extreme poverty.

  • A sole able-bodied unemployed person. As being one of the key eligibility criteria for social assistance, the unemployment requirement indicates that employment, particularly in the formal sector, and benefits receipt cannot be combined, regardless of the level of a family’s assessed income. This could potentially create work disincentives or incentives to work only informally in order not to lose the benefit. In addition, it discourages the second able-bodied family member to look for a formal job, and instead declare himself as a full-time caretaker of other dependent family members.

  • Ownership of productive assets and asset test. An ownership of property or land larger than 0.5 hectares, regardless of a family’s capacity to use it productively, disqualifies a family from receiving social assistance. In addition, it is not straightforward how the CSWs could verify all family assets, either at the time of the application or during the months that a family receives social assistance transfers. The examples of these assets are investment shares, saving bonds, saving pension, parts or full ownership rights to any property or business in Kosovo and other countries, machinery that can be used for commercial purposes, and gifts from any persons.

Figure 7.Coverage: Percent of Beneficiaries in the Bottom and Top Quintiles

Source: World Bank’s ASPIRE data.

57. MLSW has limited technical means of measuring applicant’s monetary income. As eligible families have no working members, their assessed income could come primarily from remittances, rent, as well as other benefits, transfers, and other types of pension, all of which are relatively difficult to keep track without having access to a well-linked database of beneficiaries within the MLSW departments and between the MLSW and other agencies.

58. In order to counter the lack of reliable information on family income, substantial discretion is granted to the CSWs. Without formal information on income, the CSWs rely their screening and verifying processes mainly on family and neighborhood visits. For example, families will be required to declare their income sources when the CSW officers notice a change in their durable goods—such as household appliances—during the visit. Such level of discretion may lead to incorrect exclusion of some families, as well as fraud and favoritism—especially given that CSW officers are municipal employees.

59. The equivalence scale implied by household size specific benefit levels is well below those habitually used in most advanced and emerging economies. Under the current benefit structure, the monthly amount of social assistance does not appear sufficient to bring beneficiaries out of extreme poverty. For example, at the 2011 extreme (national) poverty line, a family with four members would require about EUR 147.5 to EUR 172.4 a month. However, a four-member family’s current MGSR, together with the additional child allowance ranges only from EUR 81.3 for a family with no children to EUR 96.3 for a family with three children (Figure 8). An incremental amount of the MGSR Rate in response to an additional family member is very small and does not reflect adult equivalent family size: it is 37.5 percent for the second household member, and 12.5 percent of the single person rate thereafter until the fifteenth member.

Figure 8.MGSR vs. Food Poverty Benchmarks by Family Structure

(Total family member / Children)

Source: MLSW, and IMF staff estimates.

60. The adjustments of social assistance benefits have been nonsystematic. The Law on the Social Assistance Scheme states that the benefit level should be adjusted according to the consumer price index (prior to the 2012 legal amendments) and to the price of food basket (as a result of the 2012 legal amendments). However, the MGSR has been adjusted only three times, on an ad hoc basis: in 2003 following the adoption of Euro, additional EUR 5 in 2009, and additional 25 percent in 2015.

61. The income assessment and family size used to calculate the benefit level exclude some—but not all—types of pension and their recipients. The household definition used to determine a family’s eligibility for social assistance includes all family members living together and sharing living costs. In contrast, the household definition for calculating a family’s assessed income and benefit level omits basic pensioners and permanently disabled persons while pensions from other schemes—such as the war-invalidity pension—are included in these calculations.21 Additional inconsistency arises as pensioners receiving contributory pensions are counted in the family size but only the difference between their contributory pension and the amount of basic pension is regarded as a family’s assessed income (Table 2).

Table 2.Treatment of Pensioners and their Pension in the Social Assistance Scheme
Household definition for the calculation of
EligibilitySocial assistance benefit level
Family’s sizeFamily’s income
Basic pensionIncludedExcludedExcluded
Contributory pensionIncludedIncludedPartially included 1/
Other pensionIncludedIncludedIncluded

Only the difference between contributory and basic pension.

Only the difference between contributory and basic pension.

62. Institutional arrangements make it difficult to enforce uniform application of the regulatory framework. The CSWs, which are under their respective municipalities, are responsible for screening social assistance applications based on the information asked in the application and family and neighborhood visits. Subsequently, the MLSW carries out an additional check with the database from the tax administration offices, KPST, and other MLSW departments before making the transfers. The MLSW also conducts family visits on a regular basis—similar to and can overrule the CSWs’ screening. In addition, the appointment process at different levels of leadership of the CSWs is inconsistent. While the CSW director is appointed independently by the municipality, the head of the social assistance unit is appointed jointly by the municipality and the MLSW and cannot be removed without the Ministry’s consent.

C. Recommendations

63. Revise eligibility criteria for social assistance.

  • Discontinue pre-filtering applicants on the basis of Categories 1 and 2; alternatively

  • Increasing the age limit of the youngest child for Category 2 to 18 years;

  • Rely on asset and implied asset tests only when it is not possible to obtain accurate information on income;

  • Permit more than one Category 2 able-bodied family member to register as unemployed; and

  • Remove all employment restrictions from the list of eligibility criteria.

64. Establish a consolidate database of budget-financed cash transfers, enabling summation over families and households.

  • Consolidate into a single database government-administered benefits and recipient beneficiaries;

  • Establish an automated data exchange platform between the benefit database and the tax authority; and

  • Establish regular data exchange between the benefit database and financial service providers licensed to execute money transfers benefitting natural persons.

65. Gradually adjust social assistance benefits to reflect basic living costs and use standard equivalence scale.

  • Increase the MGSR for a single-member family at least to the level of the extreme (food) poverty line; and

  • Define the MGSR for families using equivalence scales in line with international practice.

66. Incorporate all income into the income test when establishing eligibility for and the level of social assistance.

  • Include all types of pension received by family members; and

  • Include family members aged 65 years and above, regardless of the type of benefit received, in the family’s assessed income.

67. Establish a consistent legal and regulatory framework encompassing all pension schemes.

  • Harmonize definitions and criteria, and resolve conceptual and practical inconsistencies.

68. Assess welfare and fiscal consequences of policy choices.

  • Revise laws and procedural rules concerning the legislative process to ensure that not only proposed but also adopted legislation materially impacting the financial position of state financed pension schemes is accompanied by fiscal impact analyses covering at least 25 years, presenting both cash and accrual indicators;

  • Supplement the annual budget bill with long-term projections concerning the financial position of all state financed pension schemes, including an explanation of assumptions, main drivers of change and the factors behind the changes since the previous projection; and

  • Consider introducing a rule of mandatory offsetting: it requires that when parliament votes a change to the law that implies an increase in net fiscal outlays on social welfare programs, it has to be accompanied by other measures generating the revenues necessary to offset the net increase in expenditures.

69. Indexation of benefits should be automatized.

  • Make price indexation of benefits universal and automatic;

  • Remove provisions granting cabinet discretion over the extent of benefit indexation; and

  • Establish clear rules for cases when indexation of benefits can, temporarily, be suspended.

70. Subject all pension benefits to personal income tax and to health insurance contribution.

  • Discontinue the categorical tax exemption of pension benefits; and

  • Compensate individuals receiving benefits not exceeding EUR 75 for the personal (employee) health insurance contributions.

71. Allow people benefiting from social benefits to work.

  • Consider permitting employment while in receipt of social welfare payments (pensions or social assistance); and

  • Introduce a claw-back mechanism with a relatively steep tapering factor.

V. Other Issues

72. Residency requirements are particularly weak. Residency requirements are based on local civil registration documents that are issued to all citizens of Kosovo, irrespective of their effective place of residence. Furthermore, the combination of temporary benefit suspension (6 months) and retroactive payouts (of 6 months until December 2015, now 3 months) leads to a situation where citizens can effectively return to Kosovo only once every 9 months (in the past 12 months) without forfeiting any benefits—though effectively living (and possibly benefiting from other incomes) abroad. The role of residency and citizenship in benefit entitlements warrants further study—with a need for a much clearer conceptual distinction between benefits that can be classified as social insurance (and thus contributor-based) and social assistance (residence-based).

73. The large degree of discretionary power granted to the government—on issues such as benefit indexation—raises questions of public governance. Key pension system parameters such as eligibility conditions, the rules of indexation of benefits as well as the changes in the retirement are essential parameters of a pension scheme and should not be annually decided in a discretionary way by a government. They heavily influence the long-term sustainability and adequacy of a system, and cannot be considered independently from benefit levels. The argument is further strengthened by the rather loose criteria to be met by the government in order not to apply the cost of living adjustments: the absence of budgetary options, which is clearly on endogenous parameter as spending priorities are endogenous.

74. There is no framework in place to estimate and report the medium and long-run costs and benefits of policy choices. Ex ante evaluations can guide public policy choices and allow quantification of the short, medium and long-term budgetary impact. They reduce the need for discretionary decisions by the government running counter to the spirit of the laws such as, for example, partial implementation of laws.

75. All social transfers are income tax exempt, introducing undue differentiation across various sources of income. The nontaxation of public pensions raises questions of efficiency and equity. Equal tax treatment can help alleviate inactivity traps created or at least reinforced by nontaxation of pensions. For example, the net income from contributory pensions is not too dissimilar from the net income of a minimum wage of 170 EUR gross (before income tax, health and KPST contribution). Applying the same rules of taxation to pension as to work income generally also contributes to a fairer burden sharing across the population. It improves both vertical and horizontal equity by using total income as an indicator of ability to pay, irrespectively of the specific nature of the income.

Appendix I. Basic Features of Age and Disability Benefits
Eligibility
TitleTarget groupAgeEarnings testedPublic pension testedType/Amount (EUR/month)
Basic pensionAll65-NoYesFlat (75)
Contributory pensionBeneficiaries based on law from before 1999*65 -NoYesEducation-linked (158-240)**
Disability pension100 % disabled<65Yes (Categorical)YesFlat (75)
Work disability pensionWork accident or professional disease<65Yes*** (Categorical)YesFlat (75)
Family pensionBeneficiaries based onSpousYesYesFlat (75 +20% per
(Spouse, children or parents)law from before 1999* or family of work disablede <65(Categorical)eligible child)
Trepca early pensionInvoluntary unemployed > 50% disabled50-64Yes (Categorical)YesFlat (105)
Kosovo Pension Savings TrustsAll65-NoNoPhased withdrawal (at least 150) or annuity

Notes: Effectively this corresponds to a minimum eligible career before 1999 of 15 years. The benefit range for education is not yet determined—the numbers indicate the targeted range as indicated by the authorities.

The benefit range for education is not yet determined—the numbers indicate the targeted range as indicated by the authorities.

For newly recognized cases after 1999—unspecified for others.

Notes: Effectively this corresponds to a minimum eligible career before 1999 of 15 years. The benefit range for education is not yet determined—the numbers indicate the targeted range as indicated by the authorities.

The benefit range for education is not yet determined—the numbers indicate the targeted range as indicated by the authorities.

For newly recognized cases after 1999—unspecified for others.

Appendix II. Analysis of Three Social Assistance Reform Scenarios

This note discusses how costs and savings of the social assistance reform scenarios are estimated based on the Household Budget Survey (HBS) 2014 as well as total spending and total number of beneficiary households and individuals from the MLSW’s administrative statistics.22

Structure and information of the HBS

The HBS data are collected through in-depth interviews as well as diary records of daily (food) consumption and consumption of other more-durable (nonfood) goods. The survey questions include detailed information, ranging from consumption expenditure to labor force participation and income as well as household characteristics. The reference period for labor force participation and income is the last 12 months. The HBS interviews about 200 households per month, resulting in about 2,400 households (0.7 percent of the total households in Kosovo) for the entire annual data. It uses a stratified two-stage sampling design with the Census 2011 as the frame for the first stage. The sample is then stratified by region and urban-rural.

Scenario 1: Raise Category 2’s age limit of the youngest child and/or Category 2’s number of unemployed family members.

The HBS does not allow identification of individuals who actually receive social assistance benefit, an often encountered problem in survey data. However, it contains household and individual characteristic information that can be used to identify households likely passing the eligibility criteria for each of the social assistance categories—such as the total number of family members and the numbers of family members who are working-age adults, children, pensioners, and disabled persons. The first reform scenario uses this information to identify potential Category 2 families (prior to asset and income tests). Subsequently, it estimates additional families that would pass Category 2’s eligibility criteria if the criteria on the age limit of the youngest child and the number of unemployed persons were changed. To obtain the number of additional eligible families, the percentage change in the number of eligible families based on the HBS is applied to the actual number of Category 2 families (from the administrative statistics). The scenario assumes that, on average, their income and assets are not systematically different from those who are currently receiving social assistance benefits. Then, the additional cost is equal to the estimated number of additional eligible families multiplied by the average social assistance benefits currently paid to Category 2 households.

In addition to assets and income, family size is used to determine the social assistance benefit amount. Table AII/1 and AII/2 compare the numbers of family members (who are included in the calculation of the benefit amount) and children (who would receive an additional child allowance of EUR 5 per child) of families that likely passing current Category 2’s eligibility criteria and the alternative criteria. The average family size is marginally larger for Category 2 families with more than one unemployed person, remains rather unchanged for Category 2 families with the youngest child aged between six and eight years, and declines slightly for Category 2 families with the youngest child aged 18 years. These trends suggest that, taking household size into account, the additional cost arising from lifting the age limit of the youngest child may be slightly overestimated (when changing the age limit significantly) while that due to allowing for more unemployed family members could be underestimated.

Table AII/1:Average Number of Family Members of Potential Category 2 Families
Family size 1/Maximum unemployed persons (in Category 2)
1-Base2345
Age limit of the youngest child (in Category 2)5-Base4.14.24.34.44.4
64.14.14.34.34.3
74.14.14.34.34.3
84.14.14.24.34.3
183.73.73.83.83.8

The family size presented include only family members who are included in the calculation of the benefit amount.

Source: HBS (2014)

The family size presented include only family members who are included in the calculation of the benefit amount.

Source: HBS (2014)
Table AII/2:Average Number of Children of Potential Category 2 Families
Number of childrenMaximum unemployed persons (in Category 2)
1-Base2345
Age limit of the youngest child (in Category 2)5-Base2.92.93.03.03.0
62.92.93.03.03.0
72.92.82.92.92.9
82.92.82.92.92.9
182.42.42.42.52.5
Source: HBS (2014)
Source: HBS (2014)

Scenario 2: Include pension and pensioners into a family’s assessed income and family size respectively.

Among the HBS families previously identified as potentially being in either Category 1 or 2 according to their household characteristics, the second scenario estimates whether these families would still receive any social assistance benefits if family members who are pensioners and their pension income are included in the calculation of the benefit amount. First, family members who are aged 65 and above are added to the family size used to calculate the family-specific MGSR. Second, eligible families will be excluded from social assistance benefit if their pension income exceeds the MGSR. Similar to the first scenario, the number of deductible families is obtained by applying the percentage changes in the number of eligible families based on the HBS to the actual number of Category 1 and Category 2 families respectively. By assuming that families with and without members aged above 65 years are not systematically different from each other, the additional saving of the second scenario is then equal to the estimated numbers of deductible families in each of the social assistance categories, multiplied by their respective average social assistance benefits.

For both Category 1 and Category 2 families identified in the HBS, the numbers of eligible family members and children are smaller among those families that would not receive any social assistance benefits if their old-aged members were included (Table AII/3). This implies that the deductible families under this scenario are likely to currently be receiving a smaller amount of social assistance benefit than the average. As a result, the estimated savings may be overestimated and should reflect upper-bound savings.

Table AII/3:Average Size of the Families that are Likely Affected by Scenario 2
Category 1 familiesCategory 2 families
Excl. pensioners 1/CurrentExcl. pensionersCurrent
Family size2/2.22.73.54.1
Children2.12.52.32.9

The “Excl. pensioners” column represents families of which social assistance benefits would be taken away under Scenario 2.

The family size presented include only family members who are included in the calculation of the benefit amount under the current benefit regime (i.e. excluding old-aged family members).

Source: HBS (2014).

The “Excl. pensioners” column represents families of which social assistance benefits would be taken away under Scenario 2.

The family size presented include only family members who are included in the calculation of the benefit amount under the current benefit regime (i.e. excluding old-aged family members).

Source: HBS (2014).

Scenario 3: Make the social assistance scheme become completely income-tested, and adjust the benefit level of social assistance to a standard adult equivalence scale.

The estimated impacts of the third scenario rely entirely on information from the income module of the HBS. First, an eligible family is defined as any family with formal income (including pension) below the current MGSR. Second, the total social assistance spending is equal to the sum of the difference between formal income and the MGSR of all households. Various structures of the MGSR are used in the sub-scenarios, namely the current MGSR structure, the application of an equivalence scale to the current single-person MGSR, and the application of an equivalence scale to the level of 2011 extreme (food) poverty line. Regarding the equivalence scale, the results from applying three scale options are presented: the adjusted adult equivalence scale, the Eurostat (or OECD-modified) scale, and the square-root scale. These scales are associated with largely different impacts on total spending.

Sources of estimation uncertainty

It is important to emphasize that the estimated costs/savings of the three reform scenarios rely greatly on the representativeness of the HBS sample and the (self-reported) income information. The under-representation of the poor could potentially underestimate (overestimate) the reform costs (savings), and vice versa. Likewise, the uncertainty of the estimated impact may be attributable to the discrepancies between the HBS (self-reported) income and family income information observed by the MLSW and CSWs. Furthermore, the estimation of the fiscal impacts does not take into account any behavioral responses, in particular on labor market participation.

Appendix III. Comments on The Draft MLSW Pension Concept Note

The draft concept paper prepared by the Ministry of Social Welfare and shared with the mission is an important step in taking stock of the current situation and identifying problems and policy options going forward. It provides an overview of the pension environment and its aims in Kosovo, an international perspective, as well as a reflection on desirable reforms. The document is focused on pension policy, social assistance benefits are outside of its scope.

The paper identifies three policy alternatives going forward and expresses a strong preference for one of them. First, a status quo scenario with no major changes to either laws or implementation rules and practices—considered to be undesirable. Second, a more complete recast of the legal and regulatory environment associated with profound changes to system administration. Third, more selective change of the regulatory and legislative framework coupled with a focus on specific implementation measures. The latter option is clearly identified as the preferred option.

More fundamental legal and regulatory reform is sidelined though it likely dominates the preferred incremental approach—even in the short-run. The paper unnecessarily links more fundamental legal reform to the (costly) creation of a new and separate Agency for the Management of Pensions with an extensive branch network across the territory—a proposal running counter to tendencies observed in some countries to centralize and automate services. Also, it is clear that the current legal and regulatory framework is far from self-explanatory and poses numerous problems. With the progressive implementation of new schemes, it is giving rise to an ever-increasing list of conceptual inconsistencies, legal incompatibilities and situations of regulatory vacuum that might be easier to address and administer with a new comprehensive public pension law than by a further refinement of and accommodation in the existing framework. For example, depending on the choice of an individual for veteran versus contributory benefits, eligibility benefits for family members are greatly different and hard to justify from a social welfare point of view.

The preferred reform pathway focuses on 2 key policy areas—tightening of residence requirements and means-testing of basic pension benefits. The first of these policies would effectively convert the program into a social pension scheme only accessible to people effectively living in the country—rather than the currently applied criterion of civil residence. The second policy also targets beneficiaries of these reformed social pensions by subjecting them to income-testing, including all kinds of public or private pension incomes (including KPST second pillar benefits) as well as other taxable income (mostly employment and self-employment income).

While a tighter targeting of noncontributory pensions on effective residence status is perfectly standard, the proposed income testing is less benign. As discussed in the report, there seems to be a significant discrepancy between the demographic age group 65 and older—as measured by the statistical office on the basis of the census—and the population of beneficiaries. Clearly, better administrative procedures, combined with possible minor legal changes have the potential to substantially reduce any discrepancy in this area—with substantial associated budget savings. The quoted figure in the draft concept paper of EUR 8–10 million is not implausible given the contextual information available. Public pension-testing of social pensions is likely preferable to general income or means-testing. As opposed to the note’s proposed income testing mechanism, such a mechanism would not discourage employment and self-employment of people aged 65+, and would leave incentives to formalize and save for retirement for lower income workers unaffected.

Appendix IV. Comments on The Draft Law on Police Officers’ Pension Scheme

The draft police pension law leads to a further fragmentation of the pension system. The draft presents a special pension scheme for officers of the Kosovo police force—as well as those working for the police inspectorate with police powers—further segments the population into those benefiting from preferential regimes and those left in the general scheme. The associated discrimination in benefit eligibility and levels will no doubt further reinforce other special interest groups in calling for their own special regime.

The general budget financing and earnings-linked benefits are regressive, distortive and fiscally unsound. The proposed system is a defined benefit noncontributory pension scheme, providing old-age, disability and family benefits (for spouses, children, parents and factual co-habitants) as well as early retirement benefits. Neither participants nor the police as a government entity face the economic and financial consequences of their decisions—with the fiscal burden assumed by current and future taxpayers, implying perverse redistribution between scheme members and people outside the scheme.

Career requirements are loose and regular retirement benefits are based on an excessively generous formula. A full career consists of 30 pensionable years of service, with effective years of service translating at a factor of 4/3. A full career is thus achieved after 22.5 years of work. The normal retirement age is set at 55, with a possibility for higher-ranking officers to retire up to 3 years later. By law, pensions are calculated on a final salary basis, with the last month of gross earnings serving as the basis instead of full career averages, introducing inequity and fiscal uncertainties. Regular retirement benefits are of 50 percent of the gross wage, with an increase of 2 percentage points per year of service with a maximum of 70 percent. The maximum de facto always prevails given career requirements.

Early retirement benefits are particularly unsound from an actuarial point of view. Early retirement is possible for all affiliates as early as age 50 with a reduced career requirement of 15 years of pensionable service (11.25 effective years). Benefit levels formally are of 40 percent of the gross wage increased by 1 percentage point by year of experience. A maximum of 55 percent is applied before age 55, no such maximum applies thereafter. For example, ceterus paribus, someone retiring after 22.5 years of work at 50 rather than 55 will get 5 years of benefits (at 55 percent), with no implication on post-age 55 pensions (at 70 percent).

The proposed disability and family benefits further expand the currently prevailing preferential treatment of the police. In case of disability two pensions would be available. A (work) disability pensions of between 20 and 100 percent of the last wage and an extra specific pension of 20 percent of wages paid by the police force—the latter resembling the current regime administered by the Kosovo police force itself. The proposed survivor benefits would be available irrespective of the age of the surviving spouse/parent. The amount of survivor pension is 70 percent of the police officer’s pension, with a 10 percentage point supplement for each dependent child with an overall maximum of 100 percent.

Budgetary costs are substantial in the short run. While family benefits are incompatible with other pensions financed by the state, old-age and disability pensions are not incompatible leading to a presumption of a large uptake among the (relatively young) population of pensioners. Furthermore, all envisaged benefits would also be extended to all current pensioners of the police force—leading to an increase in fiscal costs.

Fiscal risks are not limited to the short-term. The scheme proposes pay-scale indexing, de facto linking pensions in payment to wage increases of active officers/employees in the same rank as the retiree. As wages generally grow faster than prices, such indexation mechanisms generally have large medium to long term implications—rendering the generous initial replacement rates even more problematic.

Appendix V. Fiscal Impact Estimates and their Limitations

Link receipt of noncontributory pensions to continuous residency.

  • Estimated full-year fiscal impact: 0.1−0.5 percent of GDP

  • Assumptions and sources of uncertainty:

    • The estimate assumes that the entire difference between the demographically relevant population and the number of beneficiaries is explained by nonresident pensioners;

    • Population statistics based on 2011 census data which was relatively unreliable due to the lack of comprehensive system of geographic locators (addresses); and

    • Unavailability of any information or estimate of the share, distribution by age, benefit or gender of recipients residing abroad.

Limit benefit withdrawals to one noncontributory pension

  • Estimated full-year fiscal impact: negligible

  • Assumptions and sources of uncertainty:

    • The one person—one benefit rule is already mostly in place and the proposed measure will aim at limiting KSF, war invalidity and civil war invalidity pensions to be drawn in parallel with other pension benefits; and

    • No information was available about multiple benefit claims.

Make KSF and other general revenue financed earnings-related pensions contributory.

  • Estimated full-year fiscal impact: 0.1−0.3 percent of GDP

  • Assumptions and sources of uncertainty:

    • It was assumed that employee and employer contributions would be equally shared;

    • The higher end of the estimation range refers to the period when the system matures;

    • Employee contributions were assumed to remain uncompensated;

    • Employee contribution revenues are assumed to be based on current benefit expenditures;

    • No information on the age distribution of active KSF members and the expected growth of the force; and

    • No information concerning the rank structure of KSF or the wage scale.

Introduce partial disability benefit, starting at 70 percent and apply the new rules to all existing beneficiaries whose eligibility is fully or partially based on medical condition.

  • Estimated full-year fiscal impact: not available

  • Assumption and sources of uncertainty:

    • Data on beneficiary distribution by level of disability and age is unavailable;

    • Differences in original invalidity assessment is unobservable

    • There is no codified disability assessment schedule; and

    • No surveys are available concerning the health status of people with disability levels between 70 and 100 percent who currently receive no disability pension.

Make benefits income-tax liable.

  • Estimated full-year fiscal impact: 0.1 percent of GDP

  • Assumptions and sources of uncertainty:

    • It was assumed that pensioners have no other income and that all beneficiaries receive the average benefit of the relevant scheme and type of pension; and

    • No information is available about the distribution of beneficiaries’ total income.

Make benefits health-insurance liable.

  • Estimated full-year fiscal impact: 0.1 percent of GDP

  • Assumptions and sources of uncertainty:

    • It was assumed that pensioners have no other income and that all beneficiaries receive the average benefit of the relevant scheme and type of pension;

    • No information is available about the distribution of beneficiaries’ total income;

    • It was assumed that the government would fully compensate all individuals benefiting from the lowest benefits of 75 EUR (basic age pensioners, disabled, etc.) so as to leave their net income unaffected; and

    • Only personal contributions of 3.5 percent into account, as the “employer” contribution is fiscally neutral.

Harmonize family benefits across schemes and limit benefits at the level of the basic pension.

  • Estimated full-year fiscal impact: 0.2−0.4 percent of GDP

  • Assumptions and sources of uncertainty:

    • It was assumed that family pensions to working age recipients would be paid for a period of two years, at the level of the basic pension;

    • The new, general family pension and the survivors of war veterans were not considered as detailed regulations or statistics on these groups are not yet available;

    • The structure of survivors’ families is unknown; and

    • The distribution of current family benefits by recipients’ age, total benefit and gender is unavailable.

Reduce the cost of war veterans΄ pensions, through limiting total benefit receipt

  • Estimated full-year fiscal impact: 0.2 percent of GDP

  • Assumptions and sources of uncertainty:

    • Total number of beneficiaries was assumed to stay at 18,000, implying that the 12,000 people issued with a certificate of KLA-affiliation will not become eligible at a later date; and

    • The estimate is a lower bound, assuming that monthly benefit withdrawals will be limited at EUR 120 and none of the recipients will postpone drawing benefits.

Raise Category 2’s age limit of the youngest child and/or Category 2’s number of unemployed family members.

  • Estimated full-year fiscal impact: additional expenditures of 0.04 percent of GDP

  • Assumptions and sources of uncertainty:

    • All family members who appear unemployed in the HBS are assumed to register at and look for work through the employment office;

    • Additional eligible families are assumed not to be systematically different from those currently receiving social assistance benefits; and

    • No labor market behavioral response is taken into account.

Include all old-aged members and their pension in the income test.

  • Estimated full-year fiscal impact: 0.1 percent of GDP

  • Assumptions and sources of uncertainty:

    • No information is available about the (average) amount of social assistance benefits received by families with pensioners;

    • Families with and without pensioners are assumed not to be systematically different from each other; and

    • The estimated impact should reflect upper bound savings—as families with pensioners tend to be smaller than those without pensioners, and thus receive less social assistance benefits than the average household.

Make the social assistance scheme become completely income-tested, and adjust the benefit level of social assistance to a standard equivalence scale.

  • Estimated full-year fiscal impact: additional expenditures of 0.5−1.3 percent of GDP

  • Assumptions and sources of uncertainty:

    • The under-representation of poor families in the HBS could underestimate the fiscal cost;

    • Self-reported income information from the HBS is probably biased and unequal to family income information observed by the MLSW and CSWs; and

    • No labor market behavioral response is taken into account.

The legal framework consists of a variety of laws and regulations which are sometimes only partially implemented.

Otherwise benefit payments are suspended for up to 6 months. If the individual re-presents himself in person in the designated office during the period of suspension, benefit payouts are reactivated, and suspended payments are retroactively paid out for up to 3 months. In case of a no-show during the period of suspension, benefit entitlements are terminated, requiring a re-application.

Article 3 of the 2014 law on pensions financed by the state recognizes—in an unqualified way—the benefit eligibility conditions of the old system as applicable.

For example, a “parallel” partly formal partly informal education system was put in place during the period 1989-1999 by the then shadow government. The definition of “other” categories of service time still needs to be defined and refers to lower level implementation regulations, granting MLSW flexibility to substantially expanding the group of potentially eligible individuals. However, the authorities indicated that periods of involuntary unemployment by former Trepca miners would be part of this category—though the exact criteria to determine the involuntary nature of unemployment remain unclear.

This provision transforms—at least conceptually—a flat benefit into one based on length of service history.

Separate—somewhat more generous—benefits exist for the blind.

Simulation based on number of contribution pension beneficiaries in December 2015. Assumed distribution of educational attainment in the eligible population based on estimates of the authorities (as the certification process is not yet completed). Applied adjustment factors are hypothetical and applied for illustrative purposes. The increase from 140 to 158 EUR is considered as fully implemented—both measures are cumulative.

Each individual has to prove their beneficiary or eligibility status as a work disabled.

The blanket reference of the 2014 law to the applicability of the previous Yugoslav social security legislation is not straightforward and leaves numerous issues open to interpretation.

Legal uncertainty seems to persist as the law does not explicitly specify that the involuntary layoff has to have occurred in the targeted period of the 1990’s.

As indicated above, both work and involuntary inactivity periods will be fully considered towards meeting the career requirement under the category “other” sectors.

The maximum contribution rate can voluntarily is of 15 percent each. In case of self-employment, the combined rate applies.

The large community of Kosovars living and working abroad also highlights the importance of a cautious and far-sighted approach in the field of social security treaties. While some treaties and laws of the former Yugoslavia are still applied (Germany, Belgium, Macedonia, etc.), others are up for negotiation, with potentially far-reaching consequences given the size of the population living abroad. Generally, social security treaties contain amongst others rules regarding benefit to prevent situations of double contribution status, totalization rules for periods of employment, as well as situations of double benefit receipt.

The same reasoning applies to an even larger degree to spouses, children and parents of KSF, KPC and members of the police force—given their even larger generosity.

As recipients age, the difference between nonpension benefits paid to recipients of working vs. pensionable age will become more convoluted and more difficult to compare.

The Kosovo Police Force also operates a separate earnings-related disability and survivor pension scheme. Eligibility, benefit levels as well as administrative arrangements are set by the law on the police and its implementing regulations. The scheme is outside of the scope of the MLSW and is financed through the budget allocation to the police.

Social assistance applications are valid for one year for Category 1 and six months for Category 2 (see the following paragraphs on the two categories of eligible families).

Family members are dependent if they belong to one of the following groups: (i) over 18 years of age and have permanent disabilities; (ii) 65 years of age and older; (iii) full-time care-takers of person(s) in (i) and (ii) who need full-time care or of children under the age of five; (iv) up to 14 years of age; (v) between the age of 15 and 18 (inclusive) who are in full-time education; and (vi) single parent with at least one child under the age of 15.

A motor vehicle used sole for domestic purposes and not used to generate income, or a vehicle for disable persons are exempted from the asset test.

The latest poverty assessment was conducted by the World Bank in 2011. There are still no official poverty and extreme poverty lines for 2012–2015.

At the current benefit structure, a family, with members who receive other special pensions and benefits, is likely not receiving social assistance as the family income is likely be above the MGSR regardless of the family size.

Throughout the three-scenario analyses, the weight variable is taken into account.

Other Resources Citing This Publication