39146 Pension Reform and the Development of Pension Systems: An Evaluation of World Bank Assistance Background Paper Regional Summary: Europe and Central Asia Edward Palmer Director-General, Independent Evaluation Group: Vinod Thomas Director, Independent Evaluation Group, World Bank: Ajay Chhibber Senior Manager: Ali Khadr Task Manager: Emily Andrews 2007 This paper is available upon request from IEG. The World Bank Washington, D.C. ENHANCING DEVELOPMENT EFFECTIVENESS THROUGH EXCELLENCE AND INDEPENDENCE IN EVALUATION The Independent Evaluation Group is an independent unit within the World Bank Group; it reports directly to the Bank's Board of Executive Directors. IEG assesses what works, and what does not; how a borrower plans to run and maintain a project; and the lasting contribution of the Bank to a country's overall development. The goals of evaluation are to learn from experience, to provide an objective basis for assessing the results of the Bank's work, and to provide accountability in the achievement of its objectives. It also improves Bank work by identifying and disseminating the lessons learned from experience and by framing recommendations drawn from evaluation findings. IEG Working Papers are an informal series to disseminate the findings of work in progress to encourage the exchange of ideas about development effectiveness through evaluation. The findings, interpretations, and conclusions expressed here are those of the author(s) and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank cannot guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply on the part of the World Bank any judgment of the legal status of any territory or the endorsement or acceptance of such boundaries. Contact: Knowledge Programs and Evaluation Capacity Development Group (IEGKE) e-mail: eline@worldbank.org Telephone: 202-458-4497 Facsimile: 202-522-3125 http:/www.worldbank.org/ieg Acronyms and Abbreviations ADB Asian Development Bank FDC Financial defined contribution FSU Former Soviet Union FYR Former Yugoslav Republic IEG Independent Evaluation Group (formerly Operations Evaluation Department) ILO International Labour Organization IT Information Technology IMF International Monetary Fund NDC Notional defined contribution NGO Non-governmental Organization OED Operations Evaluation Department (changed its name to IEG in December 2005) PAYG Pay-as-you-go PROST Pension Reform Simulation Toolkit SIDA Swedish International Development Agency USAID United States Agency for International Development Content Preface................................................................................................................................. i 1. Introduction and Purpose of the Overview ...............................................................1 2. General Background....................................................................................................2 The Initial Years of the Transition...............................................................2 Reforming Systems for the Long-run ..........................................................7 Shortcomings in Strategies Adopted............................................................9 Adequacy of benefits .......................................................................9 Financial account schemes.............................................................10 The overall blueprint......................................................................13 Analytical capacity and long-term technical assistance.................13 3. Overview of the Evaluations of Bank Performance................................................15 Outcomes ...................................................................................................15 Bank and Borrower performance...............................................................17 Lessons.......................................................................................................18 4. Final Remarks............................................................................................................22 References.........................................................................................................................23 Tables Table 1: Key Pension System Ratios, mid-1990s .....................................................5 Table 2: Overview of Evaluation Results ....................................................................17 i Preface This paper belongs to series of 19 country and regional case studies commissioned as background research for the World Bank's Independent Evaluation Group (IEG) report "Pension Reform and the Development of Pension Systems." The findings are based on consultant missions to the country or region, interviews with government, Bank, donor, and private sector representatives involved in the pension reform, and analysis of relevant Bank and external documents. This paper was authored by Edward Palmer, Professor of Social Insurance Economics Uppsala University, Sweden and Head of Research Swedish State Social Insurance Agency, in 2005. ii 1 1. Introduction and Purpose of the Overview 1.1 In the beginning of the 1990s, the World Bank stepped up its operations dramatically in the large number of countries created with the break-up of the former Soviet Union (FSU) and the beginning of the transition of former the communist command economies in Central and Eastern Europe to democratic market economies. With the beginning of transition came the need to create institutions that supported this development. What's more the collapse of old institutions was abrupt and the need to build in their place acute. The interests of individual countries in obtaining help to create new institutions founded on experience-based knowledge coincided with the interests of the external international community in a successful transition in the region. This led to the establishment of cooperation between the World Bank and the governments of the various countries in the region to work systematically with the introduction of institutions in support of the transition. 1.2 This paper summarizes and draws conclusions from the findings of separate country evaluations of the World Bank support in the area of pension reform in countries where the Bank was a lender. The countries included in this evaluation are Albania, Bulgaria, Croatia, Hungary, Kazakhstan, the Kyrgyz Republic, Latvia, FYR Macedonia and Russia. The Bank has been involved in pension reform discussions, has helped to organize and has participated with staff in high-level discussions of reform alternatives in practically all countries in the region. Generally, the difference between Bank involvement in the countries in the Independent Evaluation Group (IEG) evaluation covered in this paper and the remaining countries is the depth of the involvement. The Bank has mobilized and provided much more technical assistance to borrowing countries, than to other countries in the region, which is explains the focus of the IEG evaluation on these countries. 1.3 The individual country evaluations are based on World Bank documents and interviews with Government officials, World Bank staff, other international institutional partners where appropriate (the ADB, the IMF, the ILO and country lending organizations such as USAID, etc.), persons outside the policy circle, such as representatives of NGO's, academic experts and journalists. In addition, the evaluations make use of existing information and analyses, including World Bank loan support documents, country strategy papers, country evaluations, economic and sector reports, etc. The individual country reports contain substantial statistical information, focusing on key demographic and pension data, which where appropriate are the data sources for this paper. 1.4 The disposition of the paper is as follows. The next section describes the general background behind the reform movement in the region and provides a brief country-by-country summary of reforms. The third section presents a summary and commentary of the evaluation of the Bank's activities in the countries reviewed. The paper ends with some general conclusions regarding the Bank's activities in the countries covered. 2 2. General Background 2.1 With independence the countries of the FSU - in the present context Kazakhstan, The Kyrgyz Republic, Latvia and Russia - inherited the pension system of the former Soviet Union. Generally speaking, the pension systems in Bulgaria and Hungary, and for that matter other countries in the region, also followed closely the FSU model. The break-up of the FSU created a number of sovereign countries, which together with the other former command economies of Eastern and Central Europe, were in acute need of assistance of creating institutions that were suitable for a market economy. This process had two phases. The first can be viewed as a period of fire-fighting, during which ad hoc measures were undertaken to create affordability in economies where national output and income sank almost overnight by around a third. The second phase was a phase of building institutions for the long-term. 2.2 This section is divided into two subsections. The first provides a brief overview of why existing systems broke down in the initial years of the 1990s, and the country strategies adopted jointly by Government and the World Bank to cope with the short-term repercussions of the collapse of economies in the initial years of transition. The second sub-section discusses in greater depth the long- term measures designed by governments in collaboration with the Bank, measures which generally supported by Bank lending projects. The Initial Years of the Transition 2.3 What all countries in the region had in common in the beginning of the 1990s was that their pension systems were confronted with extreme financial difficulties. Country after country found it impossible to meet commitments under the old regime and yearly deficits in pension budget rose to as much two percent in spite of overall employer contribution rates ­ "social taxes" just below 40 per cent. As the Bank entered the region at this time, the general picture was one of faltering economies with inappropriate institutions for supporting democracy and the market economy. Pension systems are particularly important institutions for a country because such a large part of any current population is dependent on them for their main ­ normally sole ­ source of income. In addition, it is important for workers to know in advance the extent of the state commitment. This section describes the point of departure for pension systems in the region and discusses a first wave of measures designed to put the bring pension finances under control in the early years. 2.4 Pension systems in the region existing prior to the transition were based on the pay-as-you-go (PAYG) framework. Formulas contained, typically, a fixed component and a component based on service years, and individual coefficients. Hence, there was some, albeit weak connection to actual individual working careers. Generally, throughout the region, the pension age was 55 for women and 60 for men, with a host of special rights for various groups. Service years consisted of years of employment. But, child care years counted as service years, as did conscripted military service. Normally both men and women had over 35 service years. This is important to note in the present context, because it means that there was already a foundation upon which to introduce reforms creating a 3 stronger link between contributions and benefits on a life career basis and for both genders. 2.5 In line with the principles of the old FSU system, special pension rights were awarded in lieu of better working conditions and/or higher current wages for working in unpleasant or dangerous occupations. In addition to providing service years, a special right could mean the right to receive a pension prior to the normal pension ages for men and women. They could be granted in full from age 40, if the claimant had enough years of coverage according to a specific rule for a particular occupation. For example, in Bulgaria, prior to the reform in 2000, an estimated 16 percent of pensioners were pensioners with special rights of some kind. This was more or less the general picture throughout the region. 2.6 There was no restriction on working and receiving a special pension, and most who received an early pension as a special right continued to work. Together with the low pension age, the institution of special rights contributed to the large number of pensioners. Special rights were given to people working in listed occupations such as milkmaids, bus drivers, aviation employees, the police, the military, theatre performers, wind instrument players, etc.; dwarfs; persons working in hazardous working conditions ­ e.g. in chemical, metallurgical and mining industries; and persons working in climatically extreme conditions. 2.7 There were no actuarial adjustments of early pensions, granted under special rights, or for postponement of benefit claims. Most claimed a benefit at the minimum possible age, at most 55 for women and 60 for men in all these countries, and at even much earlier ages in the case of persons qualifying for special rights. In the latter case, even though in the most extreme case a pension could be granted at the age of 40, the recipient could continue to work after claiming the benefit. 2.8 The low retirement age of women, together with the low ages at which persons with special rights qualified for a benefit led to very low system dependency ratios, in spite of rather good demographic dependency ratios. In addition, all over in the region, there was a discrepancy between the number of persons reported as employed, for example through the labor market surveys, and the number of persons contributing. This discrepancy was generally large ­ around 20-40 percent, although its size varied considerably. Table 1 summarizes the situation in four of the countries examined during the IEG evaluation. According to Table 1, in Bulgaria and Hungary, in 1995 the ratio of persons in employment to persons age 60 and over and the ratio of contributors to persons age 60 and over were very close. In Kazakhstan and the Kyrgyz Republic there was a large discrepancy, however. This difference has a lot to do with the degree of sophistication of the market economy, which in 1995 was-and still is-greater in Bulgaria and Hungary. 2.9 According to the IEG evaluations, for example, both Bulgaria and the Kyrgyz Republic developed very efficient collection administrations. The reason why compliance and coverage differ so much between these two countries is that the Bulgarian is more highly monetized, which provides a basis for collection of revenues. In the Kyrgyz Republic, low compliance is more a result of a large, very poor, traditional agricultural sector. Here there is no strong basis for collection of 4 contributions. Because of this farm areas in hectares are the basis for levying minimal taxes. In other words, country circumstances make a big difference. 2.10 Another problem affecting both contribution payments and contribution- based benefits was the wide-spread practice of paying contributions on the basis of the minimum wage. There are two sides to this. The first is that the existence of a minimum wage in this region encourages this practice. One side of this is that it may be to the short-term advantage of the system to require at least a minimum payment ­ such as that based on land area in the Kyrgyz Republic. In Bulgaria, for example, the government has tried to come to grips with this by creating minimum payment by categories, which would usually surpass the minimum wage. The other side of this practice is that the employer and the employee split the surplus resulting from payment evasion. 2.11 Changing the landscape of the pension system was another way to attempt to come to grips with the payment problem. By introducing strong links between contributions and benefits, for example, through notional defined contribution (NDC) and financial defined contribution (FDC) personal account schemes, people are penalized cutting corners. Whether and to what extent this has had any significant effect on compliance still needs to be studied. NDC scheme in 1996, accompanied with a heavy public advertising campaign to inform people of this new order. 2.12 The transition brought a collapse in both the industrial base and the agricultural sector of the transition economies. Many in the agrarian population fell more or less into the informal economy as the pre-transition command economies collapsed in the initial years of the transition. The informal economy consisted of, first, individuals working in small scale agriculture or with sporadic employment, usually in the private service sector, and, second, more or less systematic underpayment of contributions ­ perhaps based on the minimum wage as discussed above - by employers for persons in the overt labor market. It is also worth noting that during the initial years of the transition, the state as employer was among the largest delinquents, as state enterprises closed down with bankruptcy and no possibility to pay arrears. 2.13 In sum, in the early 1990s when the Bank entered the policy arena in this region, the generous rules of the pre-transition regimes were still firmly in tact, while at the same time the collapse of public enterprises and rapid spread of informality had led to a collapse in the contribution base for pension payments. Table 1 shows some key pension system indicators of four of the countries examined in the evaluation, two with relatively young populations ­ Kazakhstan and the Kyrgyz Republic-and two with older populations, more in line with Western Europe-Bulgaria and Hungary. All of the ratios of workers to pensioners are relatively low, ranging from 1.5-1.6 in Bulgaria and Hungary to 1.8 in the two central Asian countries. Once disability pensioners enter the picture, the system dependency ratios were around 1.5 in 1995. 5 Table 1: Key Pension System Ratios, mid-1990s Bulgaria Hungary Kyrgyz (1995) (1995) Kazakhstan Republic (1996) (1996) (1) Ratio of Working age Population (20-59) to Population 60 + 2.5 2.8 5.3 5.7 (2) Ratio of Persons in Employment to Population 60 + 1.9 1.8 4.3 3.7 (3) Ratio of Contributors to Population 60 + 1.7 1.8 (2.4) 2.2 (4) Ratio of Contributors to Old Age Pensioners 1.6 1.5 1.8 1.8 (5) Ratio of Contributors to Old Age and Disability Pensioners 1.5 1.5 1.5 1.6 (6) Average Old Age Benefit/ Average Wage 0.39 0.50 0.46 0.66 2.14 Together with low compliance, low pension ages and the widespread use of special rights were brought large deficits in social insurance budgets. In spite of generally too high contribution rates, revenues were not sufficient to meet payment commitments, and governments were transferring up to three to four percent of GDP, over and above the contributions collected, to prop up pension systems. State budgets were themselves in serious deficit and governments were forced to take on substantial borrowing programs. Not surprisingly, then, one of the major priorities of the time was to bring government finances into better balance, while stabilizing prices and exchange rates and creating the preconditions for positive economic growth. 2.15 To bring expenditures into line with dwindling revenues, in the first half of the 1990s, governments all over the region were forced to flatten benefits by scaling down relatively higher individual benefits to a minimum and by cancelling scheduled general indexation. Flattened benefits led to poorer conditions for pensioners, and in some cases there were even delays in payments, as governments struggled to make increasingly scarce ends meet. Relatively speaking, the decline in standards ­ with a fall in wages and a lack of employment opportunities ­ was probably as serious, if not more so, for the working population. In other words, the economic crisis in the first half of the 1990s was a general economic crisis, not just a crisis for pensioners. 2.16 By 1995, in most countries in the region, benefits no longer bore any relation to earned rights, service years or individual wage histories, where these had provided a basis for calculation of benefits prior to the period 1990-1995. Countries had de facto created flat-rate benefits. At the same time there was a general feeling that in order to create systems for the future that would be appropriate in a market economic setting it would be important to create tighter links between individual benefits and contributions. At that juncture in history this would also reward persons working in larger private enterprises and state employees, who made up a large portion of the workers who faithfully paid full contributions. A strong link would also provide an incentive for non-evasion. 6 2.17 Given the collapse of the economy and dissolution of countries' infrastructures, both in agriculture and industry, one of the great challenges in the early years was to refocus poverty alleviation so as to reach those most in need. The cut-back in pension benefits of those who were a little better off enabled a redistribution of scarce revenues to the less well off. 2.18 Consequently, in the initial phase of transition, the focus of Bank policy and advice, as well as the goals of the reform governments, was twofold: On raising the average pension age in order to reduce the number of pensioners and on creating institutional improvements that would enhance the collection of revenues. Focus was also on abolishing or phasing out special rights acquired under the old regime. 2.19 Particularly the pension age of women needed to be increased from 55 to at least 60. Life expectancy of women was generally high, and much higher than that of men. Increasing the pension age of especially women turned out to be a painful process, however, practically all countries in the region have struggled politically with this, beginning around the mid-1990s. Countries had varied degrees of success. In Hungary the pension age was raised according to a legislated schedule to 62 with relatively little protest, whereas raising the pension age of women proved impossible to discuss in Russia. In the Kyrgyz Republic, although eventually successful, an increase in the pension age came at the expense of the job of the head of the pension commission who proposed the change. 2.20 In Latvia, the minimum pension age was raised in principle ­ i.e. in the legislation ­ but the legislation was then supplemented to allow early retirement, albeit with a reduced benefit, which nullified the increase for many years. As a result women continued to retire at low ages and received extremely low benefits. As this negative result became clear with time, the Latvian government began to tighten and then abolish the early retirement rules. 2.21 Improved revenue collection was the other avenue open for improving the financial situation of countries' pension systems. This involved, among other things, improving the administration around collection services. The general advice and approach of the Bank was to consolidate collection, preferably within one institution. This enables keeping a single set of books for tax purposes, which is an advantage for employers, as is the fact that they have a single contact point for collection of all taxes and fees. From the point of view of the government, a single institution means a single set of accounts, too, and that audits, prosecution of delinquency and dealing with claims at bankruptcy are all dealt with by a single authority. The preferred institution for collection is the general revenue collection agency. 2.22 Introducing a single revenue collection agency was not uncontroversial, however, and was the subject of considerable discussion and even opposition in some countries. For example Bulgaria, the Social Insurance Administration had developed an excellent administrative track record, whereas the Bank advice, following conventional wisdom in the area, was to move collection over to the tax authority. In Latvia, transference of the responsibility for the collection of social taxes to the tax authority led to administrative confusion. During a transition period of around two years the Latvian State Social Insurance Agency 7 experienced worse services than those they were used to providing for themselves, although problems were resolved through joint efforts of the institutions involved. Of course, the change was to be addressed in the long run. In theory, in the long run there are many potential savings involved in the consolidation of tax and contribution collection with the tax authorities, which justify the short-term costs of disruption during the process of change. 2.23 In addition, in the projects covered by this review, the Bank focused on improving countries' overall pension administration by investing in structures, knowledge and practices to provide better customer services and computer technology for data processing and storage. The challenges of improving administration varied between country settings. Generally, one of the most important development paths was to devise efficient systems for collecting and storing data on contributions from employers, preferably with the help of the banking system as a payment intermediary. Another was the provision of information on rights and obligations to system participants. A third was the development of integrated services to support the development of financial account systems. 2.24 It can be argued that, ideally, in the administration of the newly created mandatory individual financial account schemes introduced in many of the countries in the region, the information and money flows could run, first through the tax authority (collection etc.), then through the social insurance agency or a separate - publicly or privately managed - clearinghouse (to keep individual investment accounts, execute fund switches and/or fund transactions in general, depending on the system design). An experience from the IEG evaluations is that countries that have introduced financial individual account schemes, e.g. Kazakhstan and Hungary, at the time of the evaluation now believe that the clearinghouse model would be the most efficient set-up. Perhaps this is a next stage in institutional development, and certainly other countries should be encouraged to consider the clearinghouse model from the outset. 2.25 In sum, the more immediate, short-term goals, pursued up into the mid- 1990s, as can be discerned from the country strategies and supporting loan documents, dealt with some of the fundamentals: the pension age, special rights and the collection of contributions and provision of benefits. At the same time, countries had been forced to flatten benefits and delay or simply ignore indexation in order to make financial ends meet. By the mid-1990s, the need for more systemic pension reform had made itself clear. Reforming Systems for the Long-run 2.26 The policy goals of countries were developed periodically through country strategies, in collaboration between the governments and the World Bank. The long-term goals of the country strategies developed involved reform of the pension systems. From the mid-1990s, Bank documents begin to focus more on structural changes in pension systems, albeit with considerable variation between countries. The long-term strategies began to be developed by reform-oriented governments, with the help of considerable technical assistance from external pension experts provided by the Bank ­ and other lending agencies (particularly 8 USAID and in Central Asia the ADB). The World Bank typically took on the role of coordinator, and often assistance was also provided by the ADB, other international organizations (the ILO and the EU) and country institutions, such as USAID. 2.27 Among the first countries to enter the arena with radical systemic reforms in the mid-1990s were Hungary, Kazakhstan, Latvia and the Kyrgyz Republic. As early as 1994-1995, the Bank was heavily involved in the initial phases of formulating reforms in Hungary and Latvia, but came in more after the fact in similar phases in the Kyrgyz Republic and Kazakhstan. The Kyrgyz Republic had already started its own process, inspired by events in Latvia and supported in the initial phase by non-Bank technical assistance. By the time the Bank had become active in Kazakhstan, the government had already decided to implement a Chile- style reform, which the Bank then supported with extensive technical assistance of various kinds. 2.28 In other countries, where Bank lending was never requested, there was nevertheless a pronounced Bank presence during this period, examples being the Czech Republic, Estonia and Lithuania, and these countries certainly benefited from technical meetings and discussions frequently arranged by the Bank. In some countries, Bank presence in the formulation of long-term strategy came later, around 1997-1998, for example in Bulgaria and Russia. The main criterion for cooperation between the Bank and the country governments was the willingness on the part of the latter, first, to implement reforms, and, second, to borrow from the Bank in order to achieve its reform goals. 2.29 In addition to technical advice, the Bank provided technical modelling assistance for evaluation of reform alternatives, primarily through dissemination of and training for using the PROST model, a model developed specifically for the purpose of analyzing pension reform. Some countries, e.g. Latvia and Russia, chose to develop their own technical models, however. In Russia, the PROST model played a role in policy formulation in 1997-98, but local modelling took over thereafter. 2.30 The reform strategies developed by governments in the region, together with the Bank, had several characteristics in common. Generally speaking, an examination of the country strategy and other documents reveals that the strategies developed had the following objectives: 1) Reduction of the risk of poverty in old age; 2) Adequate income replacement; 3) Creation of benefit formulas that link benefits closely to contributions, within the framework of a multi-pillar system, with a minimum benefit supplemented by earnings-related benefits PAYG and/or individual financial accounts; 4) The increase of minimum pension ages to levels consistent with the goals of adequate coverage in systems where benefits are tightly linked to contributions; 5) Reduction of long-tem payroll taxes; 6) Provision of incentives to lengthen working careers; 7) Increased compliance - and thereby coverage; 9 8) Reduction in the implicit debt passed on to future generations and increase fairness within and between generations; 9) Development of financial markets through the introduction of individual financial account pension schemes, thereby contributing indirectly to economic growth; 10) Establishment of viable private, voluntary pension schemes that top up public benefits; 11) Creation of an efficient administration of public benefits and contributions to both the public system and privately managed individual financial accounts, where these have been introduced; and 12) Creation of a single collector of taxes for all taxes and contributions. 2.31 Many of these objectives are interrelated. For example, a pension system that delivers an adequate benefit in old age for the average career worker will contribute significantly towards reducing the number of older persons in poverty. A minimum benefit or minimum pension guarantee for the elderly covers also persons with sporadic or short working careers, or who for some other reason are among the lifetime poor. Adequate income replacement requires affordability, which in turn requires a minimum retirement age in systems where benefits are closely linked to contributions that is set high enough so as not to allow workers to exit the labor force at too early an age to have a sufficient lifetime benefit. This is especially important for women since women can expect to outlive their spouses by seven to eight years, including the survival rates of men and women and a normal age gap between spouses. 2.32 Benefits tightly linked to contributions may also encourage compliance, and in any event penalizes non-compliance and informality. Higher minimum pension ages together with actuarially calculated benefits work in favour of postponing retirement. The introduction of defined contribution financial account schemes contributes directly to creating a tight link between contribution and benefits, while contributing to capital market deepening, which is important for a transition economy. Establishing the limits of the public system enables the establishment of supplementary private schemes. Finally, both private and public systems need to be run efficiently, with adequate services for the participants. 2.33 In summing up, it is safe to say that this package of objectives is a common thread running through all the reform strategies were the Bank was involved. Generally speaking, an overview of Bank documents shows that they all enter into the country discussions. The conditions for working with and fulfilling all these objectives vary among countries, however, and this affects the actual projects undertaken and the outcomes of Bank involvement. This is discussed in greater depth in the next section which examines the Bank's performance, beginning with outcomes in the countries covered by this evaluation. Shortcomings in Strategies Adopted Adequacy of benefits 2.34 There are at least three good reasons for discussing the adequacy of benefits. First, in any system, there will always be some lifetime poor who cannot 10 accumulate sufficient contribution-based rights to receive an adequate benefit. A second general issue involves persons with disrupted careers. In the transition setting a large percentage of the population can be expected to make the transition from the informal or semi-formal economy (perhaps from agriculture) to the formal economy during the decades immediately following reform. Whereas a new system with a tight link between contributions and benefits is desirable as a long-run pension system, in the decades following the implementation of this type of reform persons with disrupted careers, for whatever reason, risk ending up in poverty. The solution lies in creating an adequate minimum guarantee. 2.35 The second adequacy issue has to do with women. With an actuarially determined benefit, a low pension age for anyone is extremely disadvantageous. Women may find leisure at a younger pension age attractive, as well as economically feasible as long as their spouse is alive, but they will end up in poverty if they have to live on a single benefit which is small because it is claimed too early. For example, the life expectancy at age 58 of women in Kazakhstan, which is the minimum pension age, is almost 20 years. It is 55 in Russia and 58 in the Kyrgyz Republic. In contrast, it has been increased to 63 in Hungary and 62 in Latvia for both men and women. The combination of a low pension age and the actuarial financial account scheme will surely create a large number of women in poverty once their spouses have died. 2.36 There are three possible remedies for preventing female poverty in old age. The first is a higher minimum pension age of say 63-65 for women as well as men. The second is mandated unisex life tables in the computation of annuities in financial account and NDC schemes, and the third is the introduction of budget financed transfers to the financial account scheme for child care years ­ which in practice would go to women. 2.37 More generally, to date, countries have found increasing the pension age politically unpopular and, unfortunately unisex life tables have not been a topic of discussion, perhaps because of the focus on letting the market generate annuities for financial account schemes. On the other hand, the evaluations reveal that countries that have introduced financial account schemes still do not have adequate legislation and institutions for producing annuities. A conclusion is that there is still time to consider mandating unisex annuities in financial schemes. Finally, a major reason for the disrupted careers of women is childbirth and care of children in the early ages. It is a simple matter to construct child care rights that are financed with general budget revenues, and which are transferred to the individual accounts of participants. Financial account schemes 2.38 The rate of return on funds. One of the bearing arguments behind the introduction of financial account schemes is that their long-term potential rate of return is higher than the rate generated in a PAYG scheme. In Kazakhstan, for example, the real financial rate of return assumed in determining the scale of the system ­ the contribution rate ­ was an average rate of five percent. This presupposed the emergence of a domestic equity market, which has not occurred, or that funds are invested in foreign equity markets. Other countries introducing 11 financial schemes have been able to take advantage of more resilient domestic equity markets, e.g. Hungary, but even in Hungary, there is a discussion about the current rate of return. Generally, government securities cannot be expected to provide a high enough return to achieve an average real rate of return of five percent. 2.39 In the near term, which may be the next decade or more, the domestic equity market is small, undeveloped, and illiquid in countries like Kazakhstan and the Kyrgyz Republic. There are few traders, and pension funds themselves are the biggest potential customers. In a narrow market, a significant position taken in any instrument, bonds or equities, is not easily changed. However, pension funds could contribute towards stimulating new issues and trading in equities through marginal investments, which, given that they are marginal, would not significantly affect overall portfolio performance. Moreover, there is an obvious need of more investment bank/company services to screen risks and provide bundled equity investments for pension fund portfolios. In Kazakhstan, where the financial account system has been in operation since January 1998, this has still not occurred. 2.40 In sum, to achieve a high rate of return, funds need to hold foreign as well as domestic equities. This introduces an exchange rate risk and will lead to an increase in the volatility in fund returns, compared with bond dominated portfolios. Also, the domestic interest rate and exchange rate policy pursued by the governments of countries are parts of this puzzle, and have not been fully recognized in discussions of financial schemes. 2.41 Annuities. The saving accumulation scheme is to provide the single most important source of income in old age in Kazakhstan in the future and is an important component of the overall pension in Bulgaria, Croatia, Hungary, Latvia and Russia. In spite of this there is still no clear picture in these countries of what form annuity provision is going to take. In this design element, the Bank left considerable unfinished business with regard to overall design issues. The saving grace is that systems are still nascent and are either not yet in need of providing annuities or if they are the scale for any given participant is still of a negligible size ­ meaning that small lump sum payments are a feasible way to close accounts. 2.42 Note, however, that since the annuity in the financial account scheme is an important component of overall social security benefits in all countries introducing financial account schemes, as systems mature it is unreasonable to offer participants lump sum payments - or to offer or force them to take phased withdrawals with a finite contract period. This will only lead to poverty in old age for a large number of persons, namely, those who live too long. Since the state provides the safety net, eventually it will have to pick up the costs. For these reasons, it is logical to mandate that participants must be provided annuities when they convert their accumulated capital in to some form of benefit. 2.43 In Kazakhstan, for example, insurance companies are not presently participating in the market for annuities, and once they are it will be very expensive to administer annuities until they occur in a large number. The situation is similar for Bulgaria, Croatia, Hungary, Latvia and Russia. In the initial phase 12 annuities will be expensive for administrative reasons. In addition, actuaries will have to be conservative since the variance around a small number may cause losses. These two aspects mean that annuities can be expected to be unduly expensive in these nascent annuity markets. In addition, it is not clear whether participants can really benefit from possible product differentiation, other than joint life annuities, which together with a single life annuity could instead be one of the two mandated choices. In addition, there are good reasons for mandating unisex life tables in the calculation of annuities. 2.44 For all the reasons presented here, it seems that a clear an alternative to market competition for providing annuities in the mandated financial schemes is an annuity monopoly provider that provides single or joint life annuities, calculated with unisex life tables and with the whole country as the insurance group. This is the model chosen by Sweden, for example. When the insurance collective has become large and the insurance the issue of allowing competition among private insurance providers could be revisited. 2.45 Keeping Accounts: A Single Clearinghouse? Presently, in countries with financial account schemes, information is collected and stored by the individual pension funds, and each fund keeps its own accounts. This raises the question as to whether costs could be cut and information efficiency gains could be achieved by creating a single clearinghouse for the entire system, an idea that has come up in discussions on possible paths forward in the development of the systems in, for example, Bulgaria, Hungary and Kazakhstan. Latvia is the only country that has introduced a clearinghouse. 2.46 The clearinghouse is responsible for keeping accounts during the accumulation phase of participation, using one accounting method, one method of recording returns on individual accounts. This guarantees complete compatibility and, hence, comparability between pension funds, and eliminates the chance for participant's being registered at two or more funds after switches, etc. The clearinghouse can also keep accounts during the payout phase, and even be responsible for payments. 2.47 Of course, unified accounting and reporting procedures can be implemented without establishing a clearinghouse. There is even a practical reason why the formation of a clearinghouse for account keeping should be considered, at least in Kazakhstan and Hungary, where during the IEG evaluations it became clear that there is a need to invest in new software technology. For example, in Kazakhstan, the APFs are experiencing both hardware and software problems. The software, which is adapted Chilean software from pre-1998, is outdated, difficult for local IT expertise to work with, and, thus, needs to be replaced for this reason alone. The second practical reason for introducing a clearing house is that one investment instead of sixteen or more in the same technology and trained in-house staff would save on overall administration costs, which is to the clear advantage of the participants. 2.48 Financial accounts need a multi-pillar setting. A final remark is in place here. Financial account schemes are desirable, in part because they offer a better growth potential for pensions, in part because they contribute to deepening of the financial sector. The former helps to restrict the scale of contribution rates and the 13 latter helps to develop the economy ­ with positive repercussions even on the PAYG scheme. Given the pension system's potential financial volatility, it is evident that there is a need to have a stable PAYG component at the foundation of the system. Presently, a main criticism of the Kazakh reform is the absence of such a base. This weakness is generally realized by Kazakh politicians and will steer future reform policy. The overall blueprint 2.49 The challenge of reform policy is to put the pieces together to form a whole multi-pillar strategy. For example, Kazakhstan's financial account system still needs to be supplemented with some form of PAYG base that will help create adequate benefits for those who due to disruptive careers or otherwise cannot expect to receive an adequate pension. The Kyrgyz Republic provides another example of a country where the Bank and the government did not put enough effort into the overall blueprint in formulating the original legislation of 1997-98, and the situation has not changed since then. 2.50 On the other hand, a problem can arise with the overall scale of commitments. This risk arises easily with the introduction of overly ambitious financial account schemes. For example, Latvia has scheduled an increase in the contribution rate to the financial account scheme to 10 percent, with a simultaneous equivalent decrease in the NDC contribution rate. The problem is that NDC commitments already made, based on a higher contribution rate, still have to be paid in the future. This means that in the future, the contribution rate of 20 percent, split 50-50 between the two schemes will not be sufficient to pay for all past commitments. The double payment burden arises in all countries implementing financial account schemes. 2.51 Although there are country and Bank calculations of the scale of this extra cost in every country case, countries have had difficulties keeping other supporting budgetary parameters in line with what they need to be to finance the implementation of financial schemes. Countries may have counted on general budget balance, but reality may have brought large deficits that needed to be financed at the same time as the financial tier is being implemented. This puts an extra strain on the entire public budget, and if the solution chosen is to hold back on a planned expansion of the financial account scheme, the scheme (and government) may lose public trust. This became an issue in Bulgaria and Hungary in the initial phase of implementation. 2.52 Generally speaking, experience from many countries tells us that it is always difficult to reopen legislation. Lobby groups will put pressure on politicians to (re)gain special privileges. The risk is that some of the old battles will have to be re-fought. Analytical capacity and long-term technical assistance 2.53 The Bank provided technical assistance in developing analytical capacity, usually through work with the Bank's PROST model. Although PROST was adapted to Russia in the outset, in 1997-1998, the Russians eventually developed 14 their own models without the help of the Bank. Latvia also developed its own model. Generally speaking, in-house capacity was developed in every country to perform long-term technical analysis, but in most of the countries evaluated this investment was not kept alive. In the Kyrgyz Republic, there was no significant Bank presence after 1998, and as a consequence a budding analytical group never matured. The best success stories in this context are Bulgaria and Latvia where there are still trained professionals doing sophisticated analytical work. The lesson is that it takes several years to build this capacity, with continuous, but low maintenance costs to move capacity building forward over many years. 2.54 The lesson is that support in developing and maintaining technical modelling capacity is a long-term commitment. It must be built up gradually over several years. Also, the government must decide whether the home for this capacity should be maintained at the ministerial level or within the administration. Alternatively, an actuarial office or institution can be developed, as was the plan for Kazakhstan ­ however, this plan has still not become reality. 15 3. Overview of the Evaluations of Bank Performance Outcomes 3.1 The outcomes of the Bank's involvement in pension reform were judged to be at least moderately satisfactory in eight of the nine countries evaluated. In one country, Russia, the outcome was judged to be unsatisfactory. In Albania and Macedonia, the outcomes were only moderately satisfactory, however. Reading the evaluations, it becomes clear that the success of the Bank has a lot to do with correctly judging the situation in the country and adapting its role and advice thereafter. Where the outcomes were only moderately satisfactory or unsatisfactory, the Bank failed to do this, according to the evaluations. 3.2 According to the evaluations, the needs of Albania and Macedonia were far from what the Bank is normally prepared to offer, i.e. far from the rather restricted toolkit of pension reform blueprints usually rolled out. Both of these countries were at a juncture in their histories that made the level of sophisticated advice offered by Bank technical assistance difficult to transform into domestic institutional development. Albania was struggling with more basic problems of creating a market economy with a democratic foundation, with principles of rights and obligations. Macedonia was in the midst of trying to create a stable state. Neither was in a phase in their history where they were truly prepared to introduce sophisticated pension institutions, and focus should probably have been on much more mundane proposals than those discussed. Not surprisingly, the reforms in Albania and Macedonia have been judged to be unsustainable. 3.3 Russia, according to the evaluation, failed to produce a reform with logical consistency, and this, it is claimed, is where the Bank could have come in with value added. The reformers in Russia had the right principles clearly in mind, but left their "practical implementation in the grey area of political compromise." According to the evaluation, "right principles" were abandoned when the details were chiselled out. Here, there was clearly a ground for stronger Bank involvement. On the other hand, Russia had more or less shunned help from the Bank, so the Bank's opportunity to provide technical assistance had dried up around 1998, when the new government came into power. The evaluation of the Russian reform also leads to the conclusion that the reform is not sustainable. 3.4 In two countries, Kazakhstan and the Kyrgyz Republic, reform efforts were already under way when the Bank entered the picture. In Kazakhstan, the government had already decided to introduce individual financial accounts and in the Kyrgyz Republic a pension group had been working on an NDC concept, following the lead of Latvia. In these countries, when the Bank entered the process it provided technical assistance and advice designed to improve upon the already existing concepts. As a component of lending conditionality Kazakhstan was required to think about the immediate, negative poverty consequences of their proposal, but it still remains to deal with this problem in the long-run. On the other hand, the Bank and other international organizations have helped to underscore the issue of adequacy, and the government appears committed to improving upon the initial reform legislation. Similarly, the Kyrgyz NDC reform 16 is incomplete, but in this case the Bank was notably absent from the scene for a period of several years following the initial reform. 3.5 In all the six countries where the outcome ratings are at least satisfactory the Bank was able to move in directly with technical assistance and advice, and design a lending program around the country reform framework. In these countries, the Bank was more successful in building on and developing existing reform discussions. This said, it should be pointed out that in Hungary the Bank formed an alliance with the Ministry of Finance in strongly promoting the introduction of the financial account scheme, and initially it did not have the support of the community of pension experts, nor the powerful, independent Board of the Pension Fund. Nevertheless, these two camps eventually merged behind the actual reform, and ex post it is clear that there is broad acceptance for the reform in Hungary, even among experts whose opinions differed in the beginning of the reform process. 3.6 Generally speaking, where the outcomes are satisfactory a number of conditions appear to be met. These are: 1) There is a market economic and democratic foundation upon which to build institutions, and there is an absence of political chaos and strife. 2) The bureaucracy accepts and works constructively with the proposed principles, helping to formulate relevant details. 3) The Bank becomes a part of the team, providing neutral expert advice and discussing how ideas can be codified and implemented, drawing on international experience and calling in international experts. 4) A group responsible for policy analysis, with domestic experts and foreign technical advisors, is set and equipped with analytical capacity. 5) All reform efforts are coordinated in a reform group with a political mandate. 6) The principles of the reform are accepted by a broad public and political constituency. 17 Table 2: Overview of Evaluation Results Country Albania Bulgaria Croatia1 Hungary Kazakh- Kyrgyz Latvia Macedonia Russia stan Republic Evaluation Criterion Mod. Sat. Highly Sat. Sat. Sat. Sat. Sat. Moderately Moderately 1. Outcome Sat. Sat. Unsat. 2. Institutional Substantial Substantial Substantial Substantial High High Substantial Substantial Modest development 3. Unlikely Highly Highly Highly Highly Highly Likely Unlikely Unlikely Sustainability Likely Likely likely likely Likely 4. Attribution of results 4.a Bank Highly Highly performance ­ Sat. Sat. Sat. Sat. Sat. Unsat. Sat Sat. Unsat. quality at entry 4.b Bank Highly performance ­ Sat. Sat. Sat. Sat. Sat. Modest Sat. Sat. Unsat. supervision 4. c Bank Highly performance ­ Sat. Sat. Sat. Sat. Sat. Unsat. Sat. Sat. Unsat. overall 4.d Borrower Highly Highly Performance ­ Sat. Sat. Sat. Sat. Sat. Sat. Sat. Sat. Unsat. preparation 4.e Borrower Sat. Highly Sat Sat Highly Sat Sat Sat Unsat performance Sat Sat implementation 4.f Borrower Sat. Sat. Sat. Highly Sat. Sat. Highly Sat. Unsat. performance ­ Sat. Sat. compliance 4. g Borrower Sat. Highly Sat. Sat. Sat. Sat. Sat. Sat. Unsat. performance ­ Sat. overall 1Scores are absent in the IEG report and have been inferred for the purposes of this table. Bank and Borrower performance 3.7 The evaluation of Bank performance was judged as at least satisfactory in seven of nine countries. It was unsatisfactory in the Kyrgyz Republic and in Russia. The major failing is quality at entry in both countries, but also supervision in Russia. In the Kyrgyz Republic the Bank presence was neither strong prior to nor after the reform, although technical assistance was provided during the immediate reform process. The Kyrgyz reform left many loose ends, important details which were dealt with at a much higher level of sophistication in the remaining countries where Bank performance was satisfactory. The picture is similar in Russia, as has already been discussed. 3.8 Generally speaking, satisfactory Bank performance requires a considerable leadership and technical coordination investment. Where the Bank has performed satisfactorily the project manager or/and other technical experts have been a part of the domestic reform team. Being a team member puts the Bank representative in the position where there is direct insight into the daily decision process and quality control of the dozens of details that need to be gotten right to create a satisfactory reform. The comparative advantage of the Bank is experience with previous reforms and the ability to call in relevant consultants to help with the detailed legislation and institutional aspects of reform. 18 3.9 It is also noteworthy that in some of the countries evaluated there was also considerable technical assistance being provided by other agencies, e.g. the Asian Development Bank and USAID. In all cases, the evaluations reveal that the Bank took the lead in coordinating constructively the various lending and technical activities. 3.10 With the exception of Russia, perhaps most properly labelled as an unwilling client, borrower performance was satisfactory. Borrower performance was satisfactory in the Kyrgyz Republic, too, due to the dedicated work of above all the Pension Fund. The fact that borrower performance is satisfactory is not surprising since the Bank has developed lending activities primarily in countries where the government was committed to reform and help has been solicited. Lessons 3.11 There are some main lessons to be learned from the Bank's first dozen years in Eastern and Central Europe and Central Asia. This section summarizes the discussion of lessons from the separate evaluations. 3.12 The first lesson is that it is important for the Bank to be on board when important policy decisions that might require Bank assistance are being formulated. For example, the Bank entered the Kazakh discussion at a late juncture. The original Kazakh reform concept focused solely on the creation of a financial account scheme, neglecting the problems that were sure to arise due to sporadic working careers and the absence from the labor force of women in conjunction with child birth. The example of Chile, which the Kazaks were following, was known even then to provide coverage for only about two thirds of the Chilean working age population. This problem should also have been addressed and was once the Bank entered the process. Nevertheless, it is likely that the reform process and legislation in Kazakhstan would have been improved upon by earlier technical assistance from the Bank, as opposed to the stop-gap measures entered into the lending conditions. 3.13 Generally speaking, issues as significant as an overall pension reform should be easy to identify, and it should be easy for the Bank to be on board from the outset, as was the case, for example in Bulgaria, Hungary and Latvia. In these countries, the Bank worked closely with the client country throughout the process of formulating the reform. It is more difficult for all parts if the Bank enters the design discussion at too late a stage in the planning process. 3.14 A second lesson is that the training involved in policy analysis and modelling requires a major commitment from both Bank TA and the domestic counterpart. If either is absent, there will be no permanent build up in capacity. Training in this area should also be followed up with additional TA support on a less intensive basis until it is likely that the necessary level of competence has been institutionalized. The project in Bulgaria succeeded in creating such a center in the Pension Fund and in Latvia at the Ministry, but the evaluations reveal that these are the exceptions rather than the rule. This requires a long-term commitment to regularly provided technical assistance after reform implementation. 19 3.15 A third lesson appears to be that there could be a need consider how procurement procedures can be streamlined to better help the customer. The long and cumbersome procurement procedures were impediments rather than aids in at least two countries, Kazakhstan and Latvia. In fact, in Latvia, the international supplier of IT services chosen in the procurement process never delivered the contracted services, which was the major blemish of the Latvian lending project. In addition, in the latter case, the overall scale of the project was too large and would have benefited from being unbundled into small IT investment projects focused on specific needs. The Bulgarian lending project seems to have succeeded better in this respect. Seen in a larger perspective, many of the components of these lending projects are very small, and it is not obvious why a small investment project such as purchasing hard and software to support the business of ten or twenty pension funds, and perhaps a supervisory agency, should be the subject of procurement rules that better apply to large projects. More than once in the countries covered by this evaluation World Bank procurement procedures took so much time that the client was forced to find other resources to meet implementation deadlines specified in the law. 3.16 A fourth lesson is that when countries have decided to implement institutional change, the pace of implementation may go fast. This was true, for example, in Croatia, Hungary and Kazakhstan. Often, for example, the timing of the implementation may need to coincide with political windows of opportunity, which can be short. Given the lengthy institutional process usually associated with World Bank lending decisions, the Bank should be able to anticipate that project components may easily become obsolete by the time a loan is granted, in the sense that implementation process has already occurred ­ often with technical advice of Bank and other donor consultants, for example, during regular missions. This suggests that more attention should devoted to getting the timing of loans right and evaluating more closely the real need for funding given the timing of events. The general experience is that not all resources were used owing to the late arrival of funds. Nevertheless, the loans provided a liquidity margin, and perhaps a possible solution to the timing problem would be to develop alternative uses in the loan proposal. 3.17 A fifth very important lesson is that pension reform is a long process, and that the legislation process and initial year of implementation is only the beginning. More technical assistance, although probably on a scaled-down basis, is needed several years after implementation to help resolve all kinds of problems that arise ­ from public information campaigns and IT development to technical analytical help and ensuing discussions with new members of parliament who were not a part of the original process. World Bank involvement in Latvia, with the financial assistance of the Swedish International Development Agency (SIDA), was very successful in this respect. 3.18 On the other hand, in the Kyrgyz Republic, the Bank entered the process more or less after the fact, made a positive impact through influencing loan conditionality, but pulled out much too early. Bank inactivity in the Kyrgyz reform process, which was difficult politically due to legislated the increase in the pension age, contributed to leaving an unfinished reform. There was a need to work more with the details of the NDC reform and to work with the government 20 on the possible future introduction of a financial account scheme. Instead, the ensuing process led to initiatives taken by the Pension Fund, which were later criticized by the Bank ­ where continued cooperation, with technical assistance, would have been a better path, both for the Bank and the government. 3.19 A sixth lesson is that in spite of many similarities, the transition paths of Eastern European countries must differ considerably. The paths taken must reflect the specific characteristics and circumstances of a particular country. The best illustration of this, among the countries covered in this review is Albania. Here building a foundation rather than aiding in the transition was the issue. Basic values associated with citizenship as well as basic institutions had to be created. While pension reform can nowhere be considered as a mere technical problem, the Albanian experience tells us that the links between economic change and more general constitutional development should be kept more clearly in mind and more attention should be paid to the way the former may positively or negatively influence the latter. 3.20 Similarly, the absence of success in Macedonia, according to the evaluation, has more to do with the fact that at that particular stage in time, nation- building should have been and was the prime objective of the country. Although there appeared to be an urgent need to reform the pension system, the efforts devoted to this end were too sophisticated and premature. 3.21 According to the IEG country pension evaluation "if this broader attitude had been followed, the WB would have been able to anticipate the dangers of unfettered and immature financial market developments and the dire repercussions on Albanian society of the wave of unwarranted popular enthusiasm during the "pyramid period" and the inevitable betrayal that followed. The development of a more general sensibility by the Bank would be a positive element well beyond the Albanian case." 3.22 A sixth lesson, also emerging in the Albanian evaluation is that market building per se is surely not enough. A market built in vitro, the result of theoretical construction without the full awareness of governments, citizens and potential buyers and sellers may not only be ineffective but also potentially dangerous. It may ignite a crisis likely to spread more generally to society and endanger the whole transition process. More thought should be given to these general consequences and governance. 3.23 A seventh lesson is that it is important to create a strong awareness in the political arena of the "trade-offs between sacrifices and results" in the words of the IEG pension report for Latvia. The Latvians were highly aware, according to this report, and this contributed to the survival of the system when it was called into question in a referendum November 1999. Notably, according to this report, none of the many governments emerging between the implementation of the reform in 1996 to the time of the evaluation in 2004 had disowned the reform. Political ownership and awareness of the trade-offs contributes strongly to sustainability. 3.24 An eighth lesson is that projects must projects must develop an environment conducive to change and development in the upper levels of the 21 bureaucracy. This was a part of the key to success in for example Bulgaria, Hungary, Kazakhstan and Latvia, and is essential for sustainability. 3.25 A final lesson can be learned from the Kyrgyz case. Here, the Pension Fund was the strongest domestic counterpart throughout the years, with the strong support of the President's office. Given this situation, the Bank should have focused even more of its efforts on making use of this institution. In Bulgaria the situation was similar, but here the Bank succeeded in working closely with the Pension Fund, but while keeping within the overall reform group and not alienating other important counterparts. The latter happened in Hungary, where there was a split between Ministry of Finance ­ the Bank's counterpart ­ and the Pension Fund. The Pension Fund was divested of its power by the government and as a result never became a strong institutional partner in the development of the Hungarian reform. The long-term consequence is at the expense of the participants in the PAYG pillar. 22 4. Final Remarks 4.1 Once again it should be pointed out that the Bank was active in practically all the countries in the region during the period covered, providing valuable advice, arranging and participating in seminars etc. Many countries benefited considerably from the Bank's presence and engagement in pension reform, even if they did not become borrowers. Judging from the results of the countries evaluated here, the overall performance of the Bank has been satisfactory. Where the Bank has not done well, e.g. Albania and Macedonia, the major lesson is that a broader, less technical approach may be needed. The first step should be the determination of the level of advice. Perhaps a main observation is that the development of an internal team representing the various counterparts, general political consensus around the reform principles and participation of the Bank as a team member, identifying issues and bringing in the necessary expertise to deal with then, constitutes the formula for success. Finally, to succeed at the point of entry, the Bank must already have a considerable presence. For implementation to succeed the Bank must follow up its investments, including the investment in training for policy analysis. In this latter respect there is room for improvement. In the process of evaluation it became clear that even several years after the legislation had been passed and institutions had been successfully built, there was still a need for some residual help in remaining policy or design issues. Of course, in countries where reform was, in fact, incomplete, there is still much more to do. Experience shows that it may be reasonable to have a residual presence many years after the initial reform, depending on the individual country situation. 23 References Fornero, Elsa and Pier Marco Ferraresi. 2006. Pension Reform and the Development of Pension Systems: An Evaluation of the World Bank Assistance: Albania Country Study. IEG Working Paper. World Bank, Washington, D.C. _____________. 2006. Pension Reform and the Development of Pension Systems: An Evaluation of the World Bank Assistance: Latvia Country Study. IEG Working Paper. World Bank, Washington, D.C. _____________. 2006. Pension Reform and the Development of Pension Systems: An Evaluation of the World Bank Assistance: FormerYugoslav Republic of Macedonia Country Study. IEG Working Paper. World Bank, Washington, D.C. _____________. 2006. Pension Reform and the Development of Pension Systems: An Evaluation of the World Bank Assistance: Federation of Russia Country Study. IEG Working Paper. World Bank, Washington, D.C. Palmer, Edward. 2006. Pension Reform and the Development of Pension Systems: An Evaluation of the World Bank Assistance: Bulgaria Country Study. IEG Working Paper. World Bank, Washington, D.C. _____________. 2006. Pension Reform and the Development of Pension Systems: An Evaluation of the World Bank Assistance: Hungary Country Study. IEG Working Paper. World Bank, Washington, D.C. _____________. 2006. Pension Reforms and the Development of Pension Systems: An Evaluation of the World Bank Assistance: Kazakhstan Country Study. IEG Working Paper. World Bank, Washington, D.C. _____________. 2006. Pension Reforms and the Development of Pension Systems: An Evaluation of the World Bank Assistance: The Kyrgyz Republic Country Study. IEG Working Paper. World Bank, Washington, D.C.