Economic and Social Reforms for Peace and Reconciliation

Prepared by the World Bank

February 1, 2001

Table of Contents | Previous: Health

The social protection system of Kosovo faces many challenges in the coming years, as it simultaneously confronts the social aftermath of conflict, economic restructuring, the need to adjust programs and expenditures to fiscally sustainable levels, and uncertain timing of tax policy developments. This takes place in an environment where local administrative capacity requires substantial strengthening, the roles of different levels of government in social protection remain unelaborated, and local and international administrators strive to find an effective balance between the roles of the public and non-government sectors in social welfare provision.

Between the mid-1970s and 1989, Kosovo operated its own social protection system under legislation of the Kosovar autonomous provincial legislature. This was consistent with and under the general umbrella of Federal Yugoslav legislation. Cash transfers included contributory pensions and unemployment benefit, means-tested child allowances for contributors on low wages, and very tightly controlled social help for households with no resources. In 1989, Kosovar legislation was superceded and all entitlements and programs were incorporated directly into the Serbian social protection system and directed from Belgrade. New eligibility rules discriminated in several ways against ethnic Albanians. At the same time, many Albanian staff administering the various social protection programs were dismissed and replaced by Serbs. Those dismissed from their jobs in the public service also lost entitlement to unemployment benefit and child allowances and ceased to build up pension entitlements.

Between 1990 and 1999, therefore, basic social protection for many Albanian Kosovars was provided by the extended family network, by remittances from the large Albanian community outside Kosovo, and by help in kind from various humanitarian organizations. Particularly important among the latter was the Mother Teresa Society: an organized charity established in 1990 and funded by private donations from the diaspora and international organizations, with an extensive network of branch and local centers staffed by volunteers throughout Kosovo.

The Pre-Conflict Social Protection Structure ¹

The benefit structure that existed in 1999 was therefore similar to that of most SFYR, though with specific features that effectively excluded the majority of the population. Exclusionary features of the entitlement rules were exacerbated by labor market exclusion which decreased the participation of Kosovar Albanians in social insurance programs in which they would otherwise have been entitled to participate. The major benefits were:

• The pension scheme, which included age pensions (with a number of early retirement provisions), invalidity and survivor pensions, and disability allowances. In practice, only a comparatively small proportion of the population was in the scheme, almost all of them employed in the state sector. Perhaps a quarter or less of those over retirement age actually received a pension, with only 32,511 receiving an age pension as of the last payment in February 1999.² This was in part a reflection of the former Yugoslav pension system, which is purely a social insurance program, and for which participation for farmers was initiated only in the 1970s. It also reflected labor market exclusion of Kosovar Albanians, who failed to build up contribution histories in many cases after 1989. Kosovar experts also note that the system had the largest arrears of any part of the former Yugoslavia at the time of cessation of payments. The system was administered by a network of regional pension fund offices.;

• Unemployment benefit, which was both based on a contribution record and means tested. The program was administered by the Employment Service through a regional network, which also ran a range of active labor market programs;

• Child allowances were payable on a means-tested basis to people in work on low incomes. During the 1990s, eligibility rules discriminated against ethnic Albanians in two ways. First, only households with children in Serbian language schools were entitled to child allowances, effectively excluding many of the Kosovar Albanians even where parents continued to work in the formal sector. Second, limits on the number of children that could be counted for means-testing purposes worked against ethnic Albanian households, due to larger family sizes;

• Cash payments of social help were clearly seen as a last resort, carrying a high degree of stigma. These were given both on a regular basis and as one-time help for those experiencing temporary difficulties. Payments were targeted according to means and categorically, with the emphasis on those unable to work. These benefits and other social services were administered through a regional network of Centers for Social Work (CSW) similar to those found throughout former Yugoslavia. No social help payments were made for the two years prior to the conflict - when payments were last made, they went to only 39,000 people; and

• A range of social services administered by CSW, including placement of children in care; custody assessments; services for the handicapped, etc.

Financing of all aspects of the system was by contributions paid by employers and employees. In practice, the latter were largely notional. All salaries were set in terms of net take home pay. The various contributions were calculated as a percentage of this net pay, which itself remained unaffected by changes in a contribution rate. In addition, social protection was subsidized from the budget to some extent, including funds from the Fund for Undeveloped Regions, a Yugoslavia-wide fund which distributed resources from richer to poorer regions, and of which Kosovo was a major beneficiary.

Table 1 below gives estimates of staff engaged in the social protection system as of early 1999. The figures are a combination of estimates for staff actually in post, pre-1997 figures and normative values under Yugoslav legislation. For the CSW, the normative values in place previously were 1 social worker per 20,000 people; 1 lawyer and 1 educationalist per 50,000, and 1 sociologist and 1 psychologist per 100,000. However, it is important to note that most ethnic Albanian staff working in the SP system as of 1989 had been replaced during the 1990s. Many former CSW staff and central social fund management had returned to their offices to assume their former positions in 1999, though the JIAS has since carried out a more structured hiring of CSW directors and staff (see below on current system).

Following the conflict, large numbers of Kosovars lacked basic housing, furniture, clothing or cooking facilities. While self-help and donor assistance have eased the situation considerably, a significant number of households have lost the main breadwinner, and there remains significant levels of physical and psychological damage. The best indication of the overall level of need in the immediate post conflict period was probably the number of people receiving some form of help from the Mother Teresa Society: some 850,000 in September 1999 or nearly half the population. Large as this figure may seem, it was half the figure of July. When introduced in November 1999, the UNMIK social assistance system was paying benefits to 54,000 households, rising to around 80,000 in April 2000 (see below). However, as the scope of the scheme was very tightly constrained, that this may not be a realistic measure of the total number in need.

In addition to the substantial help in-kind being provided by humanitarian organizations, there is an unknown flow of funds from family members outside Kosovo, and many households receive regular stipends paid to public sector workers (who totaled around 70,000 in 2000). In addition, informal economic activity is thriving. All of this suggests that the number of people without insufficient cash income will stabilize during 2001, but will remain high for some time to come. The planned withdrawal in the near future of WFP and other donors such as UNHCR also suggests that the most vulnerable are likely to face a significant welfare shock in the first half of 2001.

Given limited budgetary resources in 1999-2000, the UNMIK administration decided - quite appropriately - that the highest welfare impact was to focus social transfer resources on an interim social assistance program. The Emergency Financial Assistance Program was administered by UNMIK with help from CSW and NGOs between November 1999 and April 2000. From May 2000, a new social assistance scheme has been implemented in stages, administered by CSW under the overall supervision of JIAS, with support from local and international NGOs (see below). The main social insurance programs - pensions, child allowance, and unemployment benefit - have not been reinstituted. Conditions that would make a restart of any social insurance programs feasible are discussed below.

The social protection function within UNMIK has core staff in place in Pristina and 5 regional centers. They work closely with CSWs throughout Kosovo, and with the Social Policy Institute. Establishing staffing levels in the local social protection system with precision remains difficult. In the 2000 budget, stipends for a total of 779 staff in the sector were provided for, of which 165 were from the Employment Service and 614 from other Social Welfare activities, the bulk of them CSW staff.⁵ A more detailed breakdown within social welfare activities was not available, but staff in pensions and child allowance centers has not been viewed as priorities, in light of the unlikely resumption of these programs in the short term. One area that is unclear, but which is of obvious importance is budgeting for staff in specialist institutions (e.g., handicapped and elderly homes; orphanages). These institutions are receiving direct donor and NGO support in several cases, but it is important that they do not "fall between the cracks" of social protection, education and health sectors.

The 2000 budget also allows for limited non-salary costs of just over DM 465,000 for the year. This will be an important item to monitor in budget execution, as administrative costs such as cash handling have historically been a significant budget item. As of the last quarter of 2000, the BPK was not charging for payments functions to JIAS, but this will need to be monitored as the situation evolves. Another critical item in rural areas is the recurrent costs of operating vehicles so that CSW staff can visit clients in outlying - particularly mountainous - areas. This is an essential function for many clients, who are unable to visit CSW due to either physical immobility or travel costs. The need for vehicles has been met to date largely through NGO donations, but operating costs are an issue, and some CSWs have already had donated vehicles stolen.

An additional issue about which there is mainly anecdotal information is the ongoing relation between the Serbian social protection system and the Kosovar Serb population. UNMIK reports that pensions have been paid since the second half of 1999 in Serb enclaves, although contributions to the Belgrade fund do not appear to have been made since the conflict. This has been done at the previous dinar rate, so that the real value of pensions is low, at around 20-30 DM equivalent per month.. It appears that none of the other payments under the Serbian system have been made, though social assistance payments within the UNMIK system have been initiated as elsewhere in Kosovo.

Most transition economies, including those in former Yugoslav republics, have confronted the need to reform inherited social protection systems. Given the likely economic situation of Kosovo in the medium to long term, it can be expected that the need for reforms in social protection will be substantial. The implications for different programs are discussed in turn below. It is stressed that the issues are in some cases medium term, in light of the delayed resumption of social insurance programs. However, broad directions need to be examined more immediately in development of a coherent Social Protection Strategy, work on which has been initiated under UNMIK with support from the World Bank and UK DFID. It will be necessary to include local experts and authorities as widely as possible in this process, in particular as new local administrations assume their responsibilities. Those same local experts see an overarching legal framework for social protection as a priority.

• The benefit structure of the pensions system is generous by Western standards, and that of many transition economies. It would be expected that any pension system that is might be introduced in the medium term would require a range of reforms which would bring it more into line with reformed transition systems throughout the region. This would include increased normal retirement ages (typically by at least 5 years for both men and women), rationalizing early retirement entitlements in order to raise average retirement ages, and flattening - perhaps entirely in the interim - pension benefits. An additional issue would be examining what are sustainable replacement rates on benefits in the medium term. Local specialists also note the important labor market role that a pension system could play in creating incentives for skilled workers to stay in Kosovo and in the formal sector;

• The inherited structure of social help payments has some unusual features by Western European standards, but the system has a track record of meeting need in a way which fits local customs and attitudes and which exercises tight control over costs. There is only limited expertise and experience in this area among the Kosovars now administering the system, and there are inevitable needs for updating of technical skills, and training in management skills that were not a feature of the previous system. Ongoing initiatives by bilaterals and NGOs (e.g., DFID, ADRA; CARE) aim to strengthen existing capacity, through improved systems and training. The social protection component of the recent Bank-financed social services project aims to build on these efforts;

• Much the same is true of child allowances. However, the feasibility of CAs in the future needs to be assessed in light of the overall social safety net, fiscal constraints, and administrative demands of multiple programs. A key strategic issue is whether any child allowance program should be focused only on those making contributions in the formal sector - as was the case previously - or made more broadly accessible, in both cases on a targeted basis; and

• The current structure and coverage of unemployment benefit are not sensible for a modern economy, and need to be considered afresh. This is a longer-term issue, however, as it is hard to see any sensible function for a contributory unemployment benefit in Kosovo over the next two to three years. The 2000 budget includes stipends for a limited number of Employment Service staff, though this is intended for active labor initiatives rather than restarting unemployment benefit payments. These efforts are supported by ILO and bilaterals. The draft 2001 budget also anticipates no unemployment benefit payments.

Sensible criteria for deciding the priority to be given the different programs in the immediate future include: operational feasibility; targeting resources on those in greatest need; and cost and affordability. For the longer term, these could be supplemented by financial sustainability; the need to take account of past obligations and expectations; maintaining a sensible structure of incentives and relativities between those in and out of work; and the advantages of some form of comparability of arrangements with neighboring countries and trade partners. Given the tendency for systems of cash payments to introduce perverse incentives or to distort the normal behavioral patterns of a society, it is also important that any desire for early action should be balanced by a degree of caution about longer-term consequences.

With respect to specific benefits, there is a range of considerations in prioritizing their reintroduction. As the social protection budget will be cash limited for the forseeable future, it will be critical to communicate to the population that gains for certain groups of beneficiaries can only be achieved at the expense of others.⁶ They are as follows:

Pension payments. There is strong pressure from various groups within Kosovo for an early resumption of pension payments, at least to those with entitlement in February 1999, and certainly resumption would go some way to help alleviate real need among potential beneficiaries. However, from a broader social protection perspective, there are powerful arguments in the circumstances immediately post-conflict for not according high priority to resumption of pensions. They include:

• The practical issues of resuming a PAYG system prior to re-introduction of a payroll tax or personal income tax are enormous;

• Payments would be badly targeted, reaching only a comparatively small proportion of those in the various vulnerable groups and without regard to immediate need. Targeting is not of course the primary objective of the pension system, but the consideration is relevant in such a constrained budgetary environment;

• The operational difficulties in identifying and locating beneficiaries can not be predicted with any confidence, and could be substantial, particularly among Serb beneficiaries;

• It would raise large equity issues with respect to ethnic Albanians who lost pension entitlements because they have been prevented from working in the formal sector during the 1990s. Working out a system which gives notional entitlements for the period of exclusion from work would probably be politically necessary, but is complicated by a range of factors, including proof of work history prior to 1989, unfunded obligations for the public pension system etc; and

• There is a complicated tangle of issues about the relative responsibilities of the Serbian and Kosovo Pension Funds – and a real possibility of disadvantaging numbers of people by too simplistic an approach to these questions.

Nonetheless, Kosovars are strongly committed to resumption of pension payments in the short term – including proposals from local experts on new financing sources - though it remains unclear to what extent this relates to all pensioners under the previous scheme, or mainly the elderly. UNMIK has been assessing options for addressing such concerns. However, the various constraints outlined above have resulted in preliminary discussions focused more on the concept of a financial supplement for the elderly, rather than payments which can be characterized as a pension strictu sensu.

Given the practical difficulties in verifying prior claims on the pension system, UNMIK is analyzing options for payments to the elderly of some form. These include: (i) paying pensions to those - or sub-sets of those - with existing entitlements, either in relation to their previous pensions or at a flat rate. This would be subject to most of the difficulties outlined above, though a flat rate scheme would avoid some problems; (ii) paying some form of flat rate "citizens pension" to all those over a certain age, irrespective of prior contributions. If such an approach were taken, it would be administratively simple and avoid many of the issues outlined above. The major risk of this approach is fiscal, with such a model potentially very costly, given increased coverage.

Social assistance: In the present circumstances of Kosovo and with the present state of public finances, it cannot be expected that cash payments to individuals will be the prime instrument for meeting need. That function is being fulfilled by humanitarian aid in kind, support from extended families within and outside Kosovo, informal economic activity, and wage payments from public sector and international organizations. In this context, there will remain an important role for the Mother Teresa network and the declining number of international agencies which will continue to function in Kosovo beyond the emergency phase. They will also participate in other services for CSW clients, though these direct services will continue to be a primary function of the public system. Nevertheless, a public social assistance scheme has been viewed as an important priority, both as a complement to humanitarian efforts, and as an indication of the continued role of the state in social protection.

The Emergency Financial Assistance Program categorically targeted those perceived to be in greatest need. The program was given high priority by UNMIK, which allocated DM 80 million for cash payments in the 2000 budget - or over 14 percent of total expenditures. The maximum monthly payment per family was DM100. The targeted social categories were: (i) the elderly over 70 years of age, who were entitled to DM 25 per month; (ii) single parent families; and (iii) families with a handicapped person. By the end of the scheme, over 80,000 payments had been made, at a total costs of DM 30 mln., of which DM 25 mln. came from the consolidated budget. Without a reliable poverty profile, it is difficult to know whether these groups are the worst off, but poverty data from neighboring countries suggests that these were reasonable choices. The most notable gap if one looks at poverty data from Albania and Macedonia is households with large numbers of children, who are significantly poorer than average in both countries. However, the fiscal burden of wide coverage among this group in Kosovo would be enormous.

The 1999 emergency program was always intended to be an interim one. As of May 2000, a new scheme has been introduced. This is takes a two tier approach. The first beneficiary category is households where there is no adult capable of work, which includes those over 65 who are not expected to look for work. The second tier will include some unemployed families, though how narrowly defined within this potentially large group is still to be worked out. The maximum benefit is DM 120 a month. The UNMIK estimated budget for 2001 for the program is DM 73 mln., with an assumption of a cash limited scheme where second category payments are likely to be quite restricted. The total caseload assumption for social assistance in 2001 is 50-60,000 families. It is clear that there will remain significant needs for refining this system going forward, in particular with respect to targeting among the second category beneficiaries. This will need to be done with a close eye on fiscal sustainability. So long as social assistance is the only public transfer program and is cash limited, the issue is somewhat less acute, but the design of the longer run scheme will need to be assessed in light of developments on pensions. If any form of additional elderly financial supplement is introduced, there would also be implications for elderly who are categorically covered under the current social assistance scheme, more so if some form of citizens’ pension was the preferred model.

Child Allowances: Social assistance payments related to household size could lead to large entitlements for those with large families – possibly larger than the amount they might earn at the bottom of the wage scale. In principle, this may result in a problem of incentives to work. This issue was addressed in the former system by child allowances. However, in the foreseeable future, it is suggested that the level of social assistance payments which will be fiscally supportable is not such as to make work disincentive problems a major issue. This is increased by the inherited stigma attached to social assistance beneficiaries. While there may be a role for child allowances in the medium term in combating child poverty, for 2001, there is no separate CA in the budget. This seems appropriate given the absence of a robust poverty profile, and the need to take decisions on specific benefits within a coherent strategy for the whole social protection system. When an individual tax system is introduced, it will be important to weigh up the competing needs for the beneficial labor incentive impact of CAs against the need to keep taxes on labor affordable, and safety net programs administratively simple and fiscally sustainable.

Unemployment Benefit. Given the likelihood of high formal sector unemployment, the absence of individual taxes, and the existence of a basic social assistance benefit, it appears that unemployment benefit payments would not be a priority over even the medium term.⁷ The experience of neighboring countries suggests that with ongoing high unemployment and limited scope for high labor taxes when introduced, there needs to be serious consideration of whether unemployment benefit is a feasible part of the future social protection system. There are of course efficiency arguments for such a program when the major issue is frictional unemployment, but it is suggested that the larger need is to stimulate growth and employment creation, and that the additional tax burden (previously up to 1.75 percent of payroll) of unemployment insurance would be difficult to justify in the circumstances.

Veterans' benefits: Since the end of the Kosovo conflict, there has been considerable pressure for a veterans' cash benefit scheme of some form, including survivor payments for those who lost their breadwinner. UNMIK have included such a scheme within the 2001 budget, on the understanding that it will be restricted in coverage, and is financed within the overall social protection ceiling, rather than being incremental. Like the social assistance program, it would therefore be cash limited. The estimated allocation for 2001 is DM 6 mln., which suggests that eligibility and benefit levels will need to be tightly controlled. The experience of neighboring countries such as Bosnia-Herzegovina and Croatia provides abundant evidence of the potential for such schemes to overwhelm social spending, and be subject to major abuses in beneficiary categorization.⁸

Over time, it would be desirable to develop a more robust profile of poverty. In order to do this, there is a strong need for reliable household level data on consumption and income, and the proportion and characteristics of households falling below various poverty thresholds. To this end, there have been estimates by NGOs and donors of a minimum consumption basket, and FAO, amongst others, has taken a lead in this area. UNMIK have also established a Working Group to review and develop poverty alleviation strategies, including household surveys and work on a poverty line.⁹ The World Bank is planning a Poverty Assessment for Kosovo which should contribute to this process.¹⁰

Generating representative and comprehensive household survey data is a priority. There have been several initiatives already, including surveys by UNFPA, IMG and KFOR. These have provided useful insights into living standards, though understandably have typically had a food security focus required for design and monitoring of humanitarian programs. To date, the most commonly used sampling frame has been the 1991 census of Kosovo. This has two major shortcomings. First, it does not reflect migration during the 1990s, and subsequent major movements of people since 1999. Second, for the ethnic Albanian population, it is based on an extrapolation of 1981 census results, due to widespread non-participation in the 1991 census.

Nonetheless, an expert assessment in April 2000 found that the basis for a representative sample of rural households existed already, and an equivalent base for urban households was constructed based on the election registration process undertaken by OSCE. This has allowed a statistically representative sampling frame to be developed. An LSMS-type survey has been designed in close consultation with a wide range of stakeholders. Pilot testing has been completed, and field work was started in early fall of 2000. The results from such a survey - which could be expected in the first half of 2001 - would provide the most comprehensive and rigorously representative overview of living standards in Kosovo to date. However, given the lead time on such an exercise, the ongoing efforts of donors and NGOs to field more focused surveys are vital.

There is no case for devolution of the financing or management of social insurance systems whenever they are reintroduced, which should be organized on a Kosovo-wide basis in the interests of creating the largest possible risk pool. The arguments are different in the case of social assistance schemes. As a longer-term system is developed, some discretion in the administration of such schemes would be desirable to avoid too rigid and broad brush an approach to different household circumstances. While there is a risk is that discretion will be exercised in ways that increase overall costs, this can be minimized by imposing local budgetary limits or by including a local element in the financing of the scheme. The role of NGOs as a tool for promoting local level accountability should also not be under-estimated. However, the lessons of neighboring countries such as Bosnia and Herzegovina need to be taken on board. In that case, complete devolution of financing and administration of social assistance has resulted in a system with high regional variations, with the poorest areas least able to provide basic benefits. Current framework regulations in Kosovo on municipal authority roles make provision for local authorities to have some responsibilities in social welfare, but the details have been left to be spelt out in regulations which will be developed after the October 2000 municipal elections. It is critical that delineation of the central and local roles take close account of the revenue base at different levels.

As noted above, the situation with Kosovar Serb population is not known in detail. While clearly local Serbs should be entitled to social assistance payments and participation in other transfer programs when they are restarted, there are a range of practical issues for the medium term about how to promote inclusion of Kosovar Serbs in the system, most particularly in areas where the local CSW is in an ethnic Albanian area. The same applies in reverse for ethnic Albanians in areas where the CSW for the municipality is located within a Serbian enclave. UNMIK have managed to address these issues to date in clearly defined enclaves. The more challenging task is where there is a dispersed minority population, and this will need to be handled in collaboration with local authorities. Again experience from Bosnia and Herzegovina is important, as the emergence of parallel delivery channels (e.g., ethnically based CSW in so-called "split cantons" in FBiH) has been a source of inefficiency and fragmentation in the social protection system. Whatever interim arrangements are made to ensure delivery of benefits to both ethnic groups must explicitly be made temporary, with a more sustainable mechanism developed as part of the overall SP Strategy.

While humanitarian aid and NGO involvement can be expected to reduce in the coming period, there is likely to remain a significant NGO presence relative to other ECA countries for the medium term in Kosovo. On the other hand, the capacity of CSW and the public system will take considerable time to strengthen. These factors, together with the continuing role of the Mother Teresa Society, suggests that developing a coherent policy on public-private complimentarities in social welfare delivery will be unusually important. There are possibilities for contracting-out and direct provision of services that are being piloted in other countries that may have important lessons for Kosovo. NGOs can also play an important role in ensuring accountability at the local level as municipalities take a stronger role in the operations of CSW.

Generation of productive employment opportunities will continue to be the major pillar of household welfare. In recognition of the importance of these issues, UNMIK have drafted a core Labor Law for Kosovo, which various stakeholders have been able to review. The draft focuses on incorporation of core ILO labor standards in Kosovo, including such key areas an non-discrimination, prohibitions on child labor etc. Given the need for rapid development of a labor relations framework, many other areas which would typically be found in other laws or regulations are also covered, including collective bargaining, collective dispute settlement, minimum wage fixing, health and safety regulations, labor inspectors and unemployment agencies.

The draft as it stands represents a significant improvement on Yugoslav labor legislation, while retaining some of the key protections of that legislation. Nonetheless, there remain several concerns on the draft which will be important to address if the new law is to facilitate job generation more effectively. It is hoped that some of these concerns – which from consultations with various stakeholders seem to be shared by local businesses, local and international NGOs, and donors – can be addressed in the approved version of the law.¹¹ Overall, promotion of small scale private sector participation in the formal economy should be a key goal of any labor legislation, and there are concerns that the current draft will either create new barriers to entry into the formal sector, or be so widely ignored as to undermine respect for labor legislation. The key areas of concern with the draft law are:

• hiring and firing provisions. There are a number of examples where the current draft is unnecessarily restrictive. These include restrictions on circumstances in which fixed term contracts can be concluded (a point of unusual important given the prevalence and uncertain time horizons of NGOs and donors); and restrictions on the circumstances in which lay-offs for economic, technological or structural reasons can take place;

• unreasonable reporting obligations on the employer, particularly with respect to recruitment and dismissals;

• wide room for unclarity, labor disputes and burdening of an inefficient court system. There are a number of key areas where the guidance given by the draft is inadequate for the practical issues confronting employers and employees. These include provisions on collective bargaining, severance pay, collective dispute settlement, powers of labor inspectors, and health and safety in the workplace. While such issues are often addressed in implementing regulations, the likely timeframe for development of these in Kosovo suggests that the core law may need greater precision; and

• the minimum wage provision. Given the economic uncertainties, unclear situation on relativities to social transfers, and underdeveloped wage bargaining institutional framework, this provision is probably, at present, best left out altogether.

• The existing social assistance system should continue as the priority cash transfer program, with joint work by UNMIK, the Institute for Social Policy, CSWs, and the World Bank/DFID-financed expert team on developing a fiscally sustainable and administratively feasible longer-term social assistance program, particularly with respect to targeting among current second tier beneficiaries;

• Work on a medium-term Social Protection Strategy should be completed as a priority, in order to help develops a coherent and sustainable structure of cash benefits and other services. This would include policy directions for reform of any inherited benefits that are likely to be restarted in the medium term, take close consideration of tax reform developments, and provide a framework for innovative public-private mixes in social welfare provision;

• Following local elections, consultation between UNMIK, local authorities, CSWs and the WB/DFID expert team would need to be initiated to consider the appropriate roles and responsibilities of different levels of government in social protection;

• Resumption of pension payments on a social insurance model should not be an early priority.

• Consistent with emerging recommendations of the Social Protection Strategy work, UNMIK should continue to assess possibilities for introduction of an elderly financial supplement, subject to resources being clearly identified;

• There is no case for reintroduction of a contributory unemployment benefit in the short to medium term;

• The reintroduction of child allowances is not an urgent priority;

• The new veterans’ scheme should remain strictly cash limited, and eligibility criteria and administrative channels developed soonest. Payments should focus on those veterans most severely disabled and survivor families, with controls on the categorization process in order to minimize the risks of abuse seen in neighboring countries;

• The LSMS should be completed and analyzed. Data and main findings should be shared as widely as possible with key stakeholders in social protection, in particular for further improvements in social assistance; and

• UNMIK should consider further revisions to the draft Labor Law (see above) which place a stronger emphasis on labor market flexibility and facilitating formalization of the gray economy.

• Based on the key recommendations of the Social Protection Strategy, work should begin on Social Protection Law which would act as a framework law for the sector;

• There should be a detailed study of the possible path to financial sustainability of the Pensions Fund over the longer term, including issues of dealing with claims of ethnic Albanians related to the 1990s. This should include a full examination of different options for the structure of the scheme, including the possibility of a phased increase in retirement age and benefit rules more in line with Western European practice;

• As it becomes possible to replace the payment of cash stipends by a proper wage and salary structure, every effort should be made to move away from the traditional approach in which salaries are denominated in terms of take-home pay and contributions are added in on top. Payments of contributory pensions should not resume without a change of this sort, even if the consequent contribution income is insufficient to match the cost of payments; and

• Further work is needed on developing approaches for ensuring that Kosovar Serb populations continue to have access to the social protection system, including assessing the implications for the CSW network.

¹¹ A Working Group of NGOs and donors has been formed to review and provide feedback to UNMIK/ILO on the draft law.