Report No. 34876 Bank Assistance to Pension Reform and the Development of Pension Systems January 5, 2006 Independent Evaluation Group Document of the World Bank Contents FOREWARD ............................................................................................................... V GLOSSARY .............................................................................................................. vi1 PREFACE ................................................................................................................ xi11 SUMMARY ............................................................................................................... XV 1. THE STRATEGY FOR PENSION REFORM ................................................ 1 The Social Protection Sector Strategy ................................................................... 3 The Structure of the Rep0rt ..................................................................................... 6 2. THE BANK’S SUPPORT FOR PENSION REFORM.................................... 9 Analytical and Advisory Services ......................................................................... 10 Lending Operations in Support of Pension Reform............................................ 12 Reviewing the Development Outcome of Pension Components ....................... 16 Summary and Conclusions ................................................................................... 17 3. QUALITY AT ENTRY FOR PENSION REFORMS ..................................... 21 When Were Only Single Pillars Considered? ...................................................... 22 Were Complementary Safety Nets Considered? ................................................. 22 Is the Economy Stable? ......................................................................................... 23 Is the Financial Sector Sound? ............................................................................. 26 Can Implementation Be Effective? ....................................................................... 29 Are Higher Saving Rates Needed t o Encourage Growth? .................................. 31 Summary and Conclusion ..................................................................................... 33 4. THE IMPACT OF PENSION REFORMS .................................................... 37 Income Security Outcomes ................................................................................... 38 Fiscal Sustainability ............................................................................................... 41 Savings and Capital Market Development ........................................................... 42 The Formalization of Labor Markets ..................................................................... 45 Development Outcomes of World Bank Assistance ........................................... 46 Summary and Conclusions ................................................................................... 47 5. BUILDING INSTITUTIONAL CAPACITY.................................................... 53 Improving the Administration of PAYG Pension Systems ................................. 54 Actuarial Forecasting ............................................................................................ 55 Improving the Regulation of Funded Pension Systems ..................................... 57 Summary and Conclusions................................................................................... 58 6. WORLD BANK COORDINATION............................................................... 61 Coordination among World Bank Groups ........................................................... 61 Cooperation with Other Donors and International Organizations..................... 63 Relationships with Clients .................................................................................... 64 Case Study Evaluation of World Bank Performance.......................................... 66 Summary and Conclusions................................................................................... 66 7 . FINDINGS AND RECOMMENDATIONS .................................................... 69 Findings .................................................................................................................. 69 Recommendations................................................................................................. 72 Boxes Box 1. Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth ........ 2 Box 2. Identifying Pension-RelatedEconomic and Sector Work .................................. 9 Box 3. How Identified Pension Projects .................................................................... 14 Box 4. Pension Reform Options Simulation Toolkit (PROST) .................................... 56 Tables Table 1. ECA and LAC Received More Support for Pension Reform Than Other Regions from Fiscal 1984 to 2005 ............................................................................. 13 Table 2. Countries with Multi-pillar Systems Received More Assistance................... 15 Table 3. Most DevelopmentOutcomes for Pension Components Were Satisfactory 17 Table 4 . In LAC. Only Some Funded Pension PortfoliosAre Well-Diversified (percentage of holdings as of December 2002) ......................................................... 38 Table 5. Real Returns Have Outpaced Wage Growth for Funded Pensions in Most Countries.................................................................................................................... 40 Table 6. Outcome ratings varied considerably, as did the reason for the rating ........ 49 Table 7 . LAC Heads the Regions in Credits and Loans Supporting Institutional Capacity Building Per Capita ..................................................................................... 53 Table 8. Economic Policy and Social Protection Sectors Had a Greater Role in Pension Reform Activities Than Other Sectors.......................................................... 62 Table 9. Many OrganizationsWork with the World Bank on Pension Reform ........... 63 Table 10. What Do Case Studies Show About World Bank Performance?.............. 66 Table 11. Even the Best Reforms Have Continuing Challenges ................................ 70 ii Figures Figure 1. What Is the World Bank's Pension Strategy?............................................... 4 Figure 2. More Sector Work Addressed Fiscal Issues and Transition Costs and Fewer Reports Discussed Public Administration .................................................................. 11 Figure 3. Many Countries Had High Inflation at Reform ............................................ 24 Figure 4 . Several Countries Had High Budget Deficits at the Time of Their Pension Reform....................................................................................................................... 25 Figure 5. Poor Financial Sectors Characterize Some ECA Multi-pillar Reformers .... 27 Figure 6 . ECA Countries Without Multi-pillar Reform Are More Likely to Have Weak Financial Sectors ....................................................................................................... 28 Figure 7. Many Reformers Had Poor Corruption Index at the Time of Reform..........30 Figure 8. Corruption Ratings Are Poor Among Some Potential Reforms .................. 31 Figure 9. Some Multi-pillar Countries Already Had High Savings Rates ...................32 Figure 10. Saving Rates Require Review Before Deciding on Multi-pillar Reform .... 33 Figure 11. Pension Funds Have Become More Diversified in Kazakhstan................39 Figure 12. Fiscal Deficits Have Grown in Many Countries with Second Pillars .........41 Figure 13. Savings Rates Increased Only in Kazakhstan .......................................... 43 Figure 14. Market CapitalizationRemains Quite Low................................................ 44 Figure 15. Post-Crisis Pension Portfolios in Argentina Fled to Government Bonds .45 Figure 16. Pension ParticipationRates Have Not Changed in LAC .......................... 46 Appendices APPENDIX A VIEWS ON PENSION REFORM: A BRIEF LITERATURE SURVEY .................................................................................................................... 75 APPENDIX B PENSION SYSTEMS IN WORLD BANK-ASSISTED COUNTRIES ............................................................................................................. 78 APPENDIX C PERFORMANCE RATINGS FOR PENSION PROJECTS................. 83 APPENDIX D MULTI-PILLAR PENSION SYSTEMS. TRANSITION COSTS. AND SAVINGS .......................................................................................................... 90 APPENDIX E INDICATOR TABLES......................................................................... 92 APPENDIX F REFERENCE TABLES ....................................................................... 100 APPENDIX G SELECTED WORLD BANK ECONOMIC AND SECTOR WORK BY REGION. (COUNTRY-SPECIFIC)..................................................................... 121 iii Attachments Attachment 1: Management Response................................................................... 129 Attachment 2: CODE Chairman's Summary ........................................................... 141 REFERENCES ........................................................................................................ 135 iv Foreword Formal pension systems constitute a n important means of reducing poverty a m o n g the aged. Inrecent years, however, pension r e f o r m has become pressing, as demographic aging, poor administration, early retirement, and unaffordable benefits have strained pension bal- ances a n d overall public finances. Pension systems have become a source of macroeconomic instability, a constraint to economic growth, a n d an ineffective and/or inequitable provider of retirement income. Inthe 1990s, the World Bank took a leading role in pension reform. The Bank’s strategy for pension r e f o r m i s formalized in Social Protec- tion Sector Strategy: From Sufefy Net to Springboard (Sfrutegy 2002), w h i c h sets forth a multi-pillar framework consisting of: ( 1)a publicly managed, tax-financed pension system; (2) a privately managed, f u n d e d scheme; a n d (3) voluntary retirement savings. Strategy fol- l o w e d up on the World Bank’s earlier policy research report, Averting the Old Age Crisis (1994), w h i c h offered a m o r e detailed prescriptive exposition of a multi-pillar pension framework. This latter report gained prominence w o r l d w i d e as favoring f u n d e d systems and pro- viding the underpinning for the Bank’s activities during the p e r i o d under review. The Bank supported a w i d e variety of pension reforms through ana- lytical and advisory services and lending operations. The Bank issued over 350 papers and publications on pension reform. The breadth of research on pensions i s impressive, covering a broad range of topics. Fiscal a n d regulatory issues, in particular, have been the focus of sub- stantial analysis. However, analyses of the living conditions of the aged tended to b e perfunctory, a n d f e w studies empirically investi- gated the l i m i t s of formal pension coverage or ways to increase it. Bank operations helped countries build institutional capacity to strengthen the administration of tax-financed pension systems and the regulation of f u n d e d pensions, providing relatively m o r e assis- tance to multi-pillar reformers. Eighty-seven percent of a l l projects with a pension component and 75 percent of the pension components themselves were satisfactory in terms of their evaluation outcome. However, case studies analyzing the longer t e r m impact of the reforms f o u n d that outcomes varied w i d e l y across countries and depended on the d e p t h of analyses, ini- t i a l conditions, institutional capacity a n d political commitment. V The Bank's strategy for a multi-pillar pension system relies on ensur- ing that sound macroeconomic policies and a n adequate financial sec- tor are in place. Incountries inw h i c h i n i t i a l multi-pillar conditions were not in place, the Bank most often supported p u r e l y PAYG re- forms that in turn contributed to fiscal objectives. Nonetheless, i na number of instances, the Bank supported multi-pillar reforms even though there were clear weaknesses i n the country's underlying eco- nomic a n d financial structure. Moreover, the Bank did not always fully consider non-contributory options to expand the social safety n e t to the populace outside the f o r m a l system. W h i l e the impact of pension reforms takes years to discern, used in- direct indicators to gauge the long-run effectiveness of the Bank's support. Bank-supported reforms have often contributed to fiscal sus- tainability. But, despite expectations, i n many countries with multi- pillar systems, funded pensions r e m a i n poorly diversified and pen- sion coverage has not increased. Also, the secondary objectives of f u n d e d pillars -to increase savings, develop capital markets, and im- prove labor flexibility - r e m a i n largely unrealized. This evaluation has several recommendations. First, to ensure well- tailored assistance t o country conditions and consistent policy prescrip- tions, the Bank needs to implement guidelines for Bank staff for the de- velopment of pension operations, p a y i n g more attention to the mini- mum macroeconomic and financial sector preconditions necessary for a multi-pillar reform. It also needs to b e careful not to oversell the bene- fits o f the secondary objectives of pension reform inits dialogue with client countries. Second, the Bank needs to ensure that client capacity to implement pension r e f o r m i s adequate, develop a checklist for capacity require- ments and provide increased assistance for building capacity. Finally, the evaluation also recommends that the Bank conduct addi- tional research on high priority issues, such as income o f the aged, the impact o f corruption and governance on the feasibility of effective pension regulation and ways in w h i c h to stimulate capital market de- velopment a n d competition. The Bank also needs to i m p r o v e internal a n d external coordination, i n c l u d i n g consensus-building among stakeholders. Vinod Thomas by Ajay Chhibber Director-General Independent Evaluation Group vi Actuarial forecasts: Forecasts used to project the long-run income and expenditure streams for pay-as-you-go (PAYG) pensions. Actuarial models can also b e developed to project income and expenditures for a variety of policy alternatives and switching patterns, including the value of the transition deficit under alternative scenarios. Annuity: A stream of payments at a specified rate, w h i c h m a y have some provision for inflation-proofing, payable until some contingency occurs, usually the death of the beneficiary or a surviving dependent. Bonosol: A pension p a i d once per year by the Bolivian government, previously called "Bolivida." It i s the first universal flat old-age pension inthe world that is n o t financed on a pay-as-you-go basis, but rather i s fully funded from a non-contributory pension fund that is invested in non-controlling equity in10 formerly state-controlled, capitalized, and privatized firms, and financed with dividend income and asset sales. Chilean pension refomz: In1981, the government gradually replaced the traditional collective PAYG system, w h i c h was managed by the state a n d w h i c h h a d defined but uncertain benefits, with a fully f u n d e d system managed by the private sector that has defined contributions but uncertain returns. M a n y countries have since implemented differ- ent versions of this reform. Contributions: Payments made by employers and/or employees to a pension system, frequently through p a y r o l l deductions; also known as a p a y r o l l tax. Coverage ratio: The number of workers actively contributing to a pub- l i c l y mandated contributory or retirement scheme, d i v i d e d by the es- timated labor force. Covered workers: Workers that are included in a formal pension p l a n (see also Coverage ratio). Defined benefit (DB): A guarantee by the pension agency or govern- m e n t that a pension will b e p a i d based on a prescribed formula, in w h i c h contributions m a y not b e tied actuarially t o benefits. Defined contribution (DC): A pension plan inw h i c h the periodic con- tribution i s prescribed a n d the benefit depends on the contribution p l u s the investment return. vii GLOSSARY Dependency ratio: The ratio of persons receiving pensions from a cer- t a i n pension scheme d i v i d e d by the number of workers contributing to the same scheme in the same period. Development outcome rating: The extent to w h i c h the project’s major relevant objectives were achieved, or are expected to b e achieved, effi- ciently. The development outcome of the pension component was iden- tified by the Independent Evaluation Group (). The development out- come for the project overall was taken from Implementation Completion Reports (self-evaluations by Bank teams), and reviews of Implementation Completion Reports (ICRs) and Project Performance Assessment Reports (PPARs). OED evaluations are independent re- views conducted by staff, frequently with the assistance of external consultants. ICR reviews are desk reviews, w h i l e PPARs are more ex- tensive and include input from client governments. Earnings ceiling: A m a x i m u m amount of earnings above w h i c h contributions to a public pension system (or multi-pillar system) are not required. Earnings-rela fed (or contribution-related) pensions: Pensions from PAYG systems that are derived u s i n g a formula related to past earnings or contributions to the system. Financial Sector Advisory Program (FSAP): A p r o g r a m of joint assess- ments by the World Bank a n d the International Monetary Fund (IMF) of the financial conditions of client countries. First pillar: A publicly managed, unfunded, defined benefit pillar; the P A Y G system (see also Multi-pillar reforms). Flat benefits: A dollar amount of pension to b e credited for each year of service or a uniformpayment to a l l entitled pensioners. Formal sector (economy): Those enterprises that fully comply with gov- ernment requirements for taxation, contributions to social insurance, a n d other legal requirements for business (see also Informal sector). Fullfinding: The accumulation of pension reserves that total 100 per- cent of the present value of a l l pension liabilities o w e d t o current members. Funded pillars (systems): Systems that are invested i n assets, i n contrast to ones that are p a i d for by taxes, either through general revenues or on a contributory basis (see also Full finding). Gross National Income (GNI): Formerly GNP, the s u m of value added by all resident producers p l u s any product taxes (less subsidies) not in- viii GLOSSARY nthe valuation of output plus n e t receipts of primary income cluded i (compensation of employees and property income) from abroad. Heavily Indebted Poor Countries (HIPC): Established i n 1996 as a joint collaboration between the World Bank and the IMF, this initiative’s a i m i s to reduce the excessive debt burdens of the world’s poorest na- tions. In1999, the initiative allowed more countries to qualify for HIPC assistance, strengthening the link between debt relief and pov- erty reduction. Implicit public pension debt (net): the value of outstanding pension claims on the public sector m i n u s accumulated pension reserves. Index o f control of corruption: The World Bank developed a comprehen- sive set of governance indicators for the anti-corruption project, Gov- ernance Matters (see Kaufman, Kraay, Mastruzzi 2004). This category measures perceptions of corruption, conventionally defined as the ex- ercise of public power for private gain. The particular aspect of cor- r u p t i o n measured ranges from the frequency of ”additional payments to get things done,” to the effects of corruption on the business envi- ronment, to measuring ”grand corruption” i nthe political arena or in the tendency of elite forms to engage i n”state capture.” Informal sector (economy): Enterprises that do not fully comply with government requirements for taxation, contributions to social insur- ances, and other legal requirements for businesses, or firms and those workers that are not included in such requirements (see also Formal sector). Legal retirement uge: The normal retirement age written into pension statutes. Mandatory pension system: A pension system for w h i c h contributions n a country or for workers i are required for a l l workers i n particular covered sectors of the economy. Market capitalization: The share price times the number of shares out- standing. Listed domestic companies are the domestically incorpo- rated companies listed o n the country’s stock exchanges at the e n d of the year. Means-tested benefits: Benefits that are targeted to the poor based on income a n d assets. Minimum confrz‘buto y period The minimum length of t i m e that contri- butions m u s t b e made t o a public pension system to receive a pension a t retirement. ix GLOSSARY Minimum pension guarantee: A guarantee p r o v i d e d by the government to bring pensions to some minimumlevel. Multi-pillar reform (system): Pension r e f o r m (system) with a first pillar that i s public (generally PAYG); a second pillar that i s mandatory and funded; and a third pillar that i s voluntary and f u n d e d (see also F u n k d pillars). Inthis report, multi-pillar r e f o r m i s used to describe nimplementing a mandatory, any r e f o r m that involves or assists i funded pillar. Normal retirement age: The usual age at w h i c h employees become eli- gible for occupational pension benefits, excluding early retirement provisions. Net official aid: Aid flows (net of repayments) from official donors to countries and territories in Part I1of the Development Assistance Committee (DAC) l i s t of recipients: including m o r e advanced Central a n d Eastern European countries, the countries of the former Soviet Union, a n d certain advanced developing countries a n d territories. Of- ficial a i d i s p r o v i d e d under terms and conditions similar to those for other development assistance. Data are in current U.S. dollars. Net oficial development assistance: Disbursements of loans made on con- cessional terms (net o f repayments of principal) and grants by official agencies of the members of DAC, by multilateral institutions, and by non-DAC countries to promote economic development and welfare in countries and territories inpart I of the DAC list of recipients. It in- cludes loans with a grant element o f at least 25 percent (calculated at a discount rate of 10 percent). Notional defined contribution (NDC): Resembles a defined contribution system inh a v i n g i n d i v i d u a l accounts that "accumulate" a l l the con- tributions of a worker, and then converting that s u m into an annuity a t retirement, but in w h i c h the r e t u r n to contributions i s 'notional', that is, i t i s not based on marketable investments in physical or finan- cial assets. Occupational pension scheme: An arrangement by an employer to pro- v i d e retirement benefits to employees. Parametric reform: A type of pension r e f o r m that maintains the struc- ture a n d administration of the system but changes k e y elements of the parameters, such as contribution rates, retirement ages, or average benefit levels. Pay-as-you-go (PAYG): Ini t s strictest sense, a m e t h o d of financing whereby current outlays on pension benefits are p a i d out of current revenues, often f u n d e d from a p a y r o l l tax. X GLOSSARY Pension coverage (see Coverage ratio). Pension system balance: The difference between pension fund revenues a n d pension fund expenditures i n a PAYG system. Pologne, H o n p Assistance a la Reconstruction Economique (PHARE): An instrument financed by the European U n i o n (EU) to assist the applicant countries of Central and Eastern Europe intheir preparations for join- ing the EU. Now that 8 out of 10 Central and Eastern European coun- tries, w h i c h previously were eligible for the PHARE program, are EU Member States. Since M a y 2004, the program has been substantially reduced and currently only Romania and Bulgaria remain PHARE re- cipient countries. Point systems: P A Y G pension systems inw h i c h pensions are deter- nw h i c h the individual's earn- m i n e d according to a "point" formula i ings are compared to the average wage. Policy and Human Resources Development (PHRD) Fund A collaborative effort between the Japanese Government and the World Bank, a n d currently one of the Bank's largest source of grant funds available to developing countries. The Fund, established in 1990, a n d i t s predeces- n1987, have p r o v i d e d sor, the Japan Grant Facility, established i nearly 2,000 grants in support of technical assistance activities to more h t an 120 countries. Pre-funding: The accumulation of contributions in a f u n d e d system. Privately managed: Funded pensions invested in assets by private pen- sion funds or private asset managers (not managed by the govern- ment). Providentfund: A fully funded, defined contribution scheme inw h i c h funds are operated a n d generally managed by the public sector. Prudent investor rule: The r u l e requires investors ( 1 )to b e p r u d e n t and act as other careful investment professionals would, (2) to diversify and thereby minimize risk, (3) to monitor their investments a n d make necessary changes, (4) to b e l o y a l and act solely in the best interests of their beneficiaries. Replacement rate: The value of a pension as a proportion of a worker's wage during some base period, such as the last year or two before re- tirement or the entire lifetime average wage. It also denotes the aver- age pension of a group of pensioners as a p r o p o r t i o n of the average wage o f the group. Second pillar: A funded, defined contribution pillar with no redistribu- tion. With more ambiguous systems (e.g., systems that are partially xi GLOSSARY funded or are managed by a tripartite board), this report classifies a system as h a v i n g a second pillar if the funds are separate from the budget and invested in assets (see also Multi-pillar reforms). Such sys- tems generally rely on i n d i v i d u a l accounts. Social pensions: Non-contributory pensions p a i d to the aged over a cer- tain age who are not receiving contributory pensions or whose con- tributory pensions are less than the social pension. Systemic pension reform: A type of pension r e f o r m that replaces the ex- isting PAYG system with a multi-pillar or other type of pension sys- t e m that diversifies the structure of benefits, administration, a n d funding of the pension system. Technical Assistance to the Commonwealth of Inkpendent States (TAUS): The m a i n EU initiative to help the countries of the Former Soviet Un- ion with the transition to a market economy. It supports democracy and the exchange o f knowledge and expertise through partnerships, links, a n d networks a t a l l levels of society and i s based on close coop- eration and exchanging experience between partners. Third pillar: A voluntary, privately managed pension pillar (see also Multi-pillar reforms). Transition cost: The cost to the government of transforming a PAYG system to a multi-pillar system, w h i c h involves rnaking the i m p l i c i t pension liability explicit. Transitional deficit: The government deficit caused by the transition cost (see Transition cost). Uncovered workers: Workers that are not included in a f o r m a l pension p l a n (see also Covered workers). xii Preface This evaluation presents an independent assessment of the World Bank’s support for pension reform activities focusing o n the period f r o m 1994-2004. This report is the first comprehensive Independent Evaluation Group () evaluation of the Bank’s involvement inpension reform, assessing the implementation of the Bank’s strategy and the re- sulting development outcomes. Since 1984, the Bank has assisted 68 countries with reform of their pension systems through more than 200 loans and credits. Inaddition, the Bank has issued over 350 papers and publications on pension reform. This report analyzes the Bank’s assistance to support pension reform to determine whether the Bank’s strategy was relevant and whether it was followed. The evaluation assesses whether pension reform decisions re- flected best practice guidelines at entry, and whether Bank-assisted re- forms achieved their social, macroeconomic, and financial objectives. The report also evaluates the Bank’s assistance inbuildinginstitutional ca- pacity, coordinating within the Bank, and cooperating with other inter- national organizations. Finally, the evaluation summarizes these findings and presents specific recommendations for the future. The evaluation i s based on a compilation developed by of all Bank lending a n d economic and sector work on pension reform, case stud- ies for 16 countries, Financial Sector Advisory Program (FSAP) as- sessments, economic indicators, desk reviews, a n d interviews with Bank staff a n d external organizations. The report was circulated to Bank management i n v o l v e d in pension reform, the Human Development Network, and the Operations Policy a n d Country Services Department. The country case studies were also distributed to the relevant Country Directors. Emily S. Andrews authored the evaluation with Debby Kim Wang (CR). Rafael Rofman (LCSHD), Ed Palmer, Elsa Fornero, Pier M a r c o Ferraresi, Jorge San Martino, Salvador Valdes-Prieto, John Piggott, a n d Hazel Bateman (consultants) authored the case studies. H.Joan M o n g a l (OEDCR) p r o v i d e d administrative assistance. The evaluation also benefited from the comments of three peer reviewers: Roberto Rocha (Office of Policy Development), John Johnson (CR), a n d L a r r y Thompson (consultant) and six external reviewers: Gary Burtless, L a u r i Leppik, E. Philip Davis, August0 Iglesias, Emmanuel Reynaud, a n d N o r i y u k i Takayama. xiii Summary 1. Pension reform is a focus of World Bank activities because pen- sions are an important part of the social safety net for workers covered by the formal pension system i nmany client countries, providing a mechanism to reduce the risks of old-age poverty a n d a means to smooth lifetime income, maintaining living standards in retirement. 2. Pension systems m u s t be fiscally and politically sustainable to achieve their income-support objective. Unsustainable pension systems can b e an obstacle to fiscal stability economic growth and poverty re- duction. Over the past two decades, the need for pension r e f o r m has become m o r e pressing i n client countries because demographic aging a n d the mismanagement of pension systems have put a strain on gov- ernment budgets, w h i c h threatens to undermine macroeconomic sta- bility a n d retirement income security. 3. Countries with high coverage rates a n d a n increasingly high percentage of the population reaching retirement age are most likely to face severe future fiscal imbalances. Countries i n the Bank’s Europe a n d Central Asia Region (ECA) are p r i m e examples. E v e n countries with l o w e r coverage and younger populations, including countries i n the L a t i n America and Caribbean Region (LAC), face fiscal issues similar to those with serious demographic problems, particularly w h e n employment i n the covered sector i s declining relative to an in- Demographic creasing number of retirees. Incountries in other regions, pension re- aging may lead form, to date, has been less of a priority. to severe fiscal imbalances in 4. This report i s the first comprehensive independent evaluation of the Bank’s involvement in pension reform. It assesses the Bank’s the future pension r e f o r m strategy a n d the resulting development outcomes for Bank assistance between 1984 and 2004. During this period, the Bank assisted 68 countries with r e f o r m of their pension systems with m o r e than 200 loans and credits. Inaddition, the Bank issued more than 350 papers a n d publications on pension reform. 5. This report analyzes the Bank’s assistance to determine whether the Bank’s strategy was relevant a n d whether i t was fol- lowed. M o r e specifically, the evaluation assesses whether pension re- form operations reflected best-practice guidelines a t entry, a n d whether the reforms achieved their social, macroeconomic, and finan- cial objectives. The report also evaluates the Bank‘s assistance i n building institutional capacity, as w e l l as additional factors that c o u l d xv SUMMARY affect r e f o r m outcomes. Finally, the evaluation summarizes these findings a n d presents specific recommendations for going forward. The Strategy for Pension Reform 6. The Bank’s strategy on pension reform i s presented in Social Protection Sector Strategy: From Safety Net to Springboard (2001) - hereafter called Strategy. Because, this document i s the only official Board-approved strategy for pension reform, it forms the basis of the evaluation. Of course, many Bank publications have influenced the di- rection of Bank assistance, inparticular, the Bank’s earlier publication Averting the Old Age Crisis (1994) -hereafter called Averting. Ineffect, Strategy ratified many of the precepts established inAverting. 7. Strategy recommends the establishment of flexible multi-pillar pension systems, consisting of three ”pillars” based on different forms of income support, as long as proper initial conditions prevail. The first pillar consists of a publicly managed, u n f u n d e d plan; the second pillar i s a mandatory, privately funded plan; and the third pillar i s a voluntary, privately f u n d e d plan. It also recommends complementary retirement income provisions for uncovered workers and the poor. 8. Based on the Bank’s strategy, the Independent Evaluation G r o u p () used the following criteria to judge soundness of pension re- forms: (1)impact on the income of the aged, (2) the nature of the fiscal policy and financial sector environment, (3) the capacity of the admin- istrative structure to operate a multi-pillar system, a n d (4) the sound- ness of regulatory a n d supervisory arrangements. The report evalu- ates the extent to w h i c h these criteria have been m e t based on statistical indicators, assessments from Implementation Completion Reports (ICRs) and Project Performance Assessment Reports (PPARs), assessments from the Financial Sector Advisory Program (FSAP), in- terviews with Bank staff a n d external stakeholders, a n d 16 case stud- ies. The Bank’s Support for Pension Reform 9. The World Bank supported a variety of pension systems, both u n f u n d e d a n d funded, through l e n d i n g operations and analytical a n d advisory services, including economic and sector work, policy dia- logue, training, and dissemination. The Bank p r o v i d e d $5.4 billion in pension-specific lending f r o m 1984 to 2005, with m o r e t han half of it issued during the fiscal 1998-2001 period. Of the countries receiving Bank support for pension reform, ECA and LAC dominated with a combined total of 40 countries. xvi SUMMARY 10. The Bank’s papers and publications on pensions provide a substantial foundation for the Bank‘s operations, policy dialogue a n d overall approach on pensions. The breadth o f analytical work i s con- siderable, with a preponderance of studies on ECA countries followed by LAC, paralleling the pattern of lending. Fiscal and regulatory issues have been the focus of substantial analysis because fiscal imbalance has been the leading reason for countries t o undertake pension r e f o r m a n d seek the Bank’s assistance. The Bank has undertaken numerous studies on complicated technical issues such as the regulation of funded pensions and the administrative costs of the f u n d e d pillars. Bank thinking on pensions has evolved over time reflecting broader discussion and accumulated country experience. 11. While the Bank’s analytical contributions represent a critical expansion o f knowledge on pension reform, economic and sector work often failed to provide country-specific guidance to assist i nproject de- velopment. For example, while the Bank has conducted poverty as- sessments i nmany countries, all too f e w offer a detailed profile of the living conditions of the aged. Similarly, while low pension coverage i s frequently mentioned as a problem, little empirical research has been conducted to identrfy policies that encourage its growth. Additionally, studies to improve public pension administration have been under- represented i n Bank work, despite the importance of p r o g r a m imple- mentation for P A Y G and mandatory, f u n d e d pensions alike. Eco- nomic and sector work has been l i m i t e d in a number o f other areas, as well, including disability a n d survivor’s pensions a n d the political economy aspects of reform. 12. While the Bank‘s lending operations have helped r e f o r m m a n y publicly managed, u n f u n d e d plans, the Bank has p r o v i d e d greater re- sources to countries developing multi-pillar systems. I nLAC, the Bank p r o v i d e d lending support for mandatory, privately f u n d e d pil- lars, which, i n one form or another, were implemented i n most coun- tries where the Bank engaged i nECA, the Bank also un- n dialogue. I dertook operations to assist the development of mandatory, privately na number o f these, however, reforms were slow in f u n d e d pillars. I coming or never implemented. With the exception of a l i m i t e d num- ber of countries that offered social pensions, particularly i n ECA, the Bank p r o v i d e d little support to develop social assistance for the aged poor, even though this was a stated element of the Bank’s strategy on pension reform. 13. Inits financial support of multi-pillar systems, the Bank has n o t taken a one-size-fits-all approach. Systems supported by Bank lending varied i n size and design. However, i t i s difficult to document whether this variation was the result of the Bank’s taking into account specific country conditions, the task team’s preference for a particular structure, or the country’s preference for a particular reform. xvii SUMMARY 14. Over three-quarters of the Bank's ratings for pension compo- nents, and projects that included pension components, h a d satisfac- tory outcomes. Based on project ratings, the Bank appears to have been more successful i nits pension reform activities inLAC than in ECA. 15. While satisfactory ratings for i n d i v i d u a l l o a n activities are im- portant, the success or failure of an operation m a y not correspond to the success or failure of a Bank pension program overall. case studies o f particular counties, w h i c h analyze the Bank's assistance to pension r e f o r m comprehensively over time, indicate that development out- comes depend on m u l t i p l e factors. Inparticular, successful outcomes depend on attention to initial conditions, effective institutional capac- ity, a n d political commitment to the reform. Was Bank Support Consistent with Initial Conditions? 16. While the primary objectives of pension reform are to reduce old-age poverty and smooth lifetime consumption, Strategy indicates that additional objectives of multi-pillar reforms m a y also b e achieved, including greater worker participation in the pension system and higher economic growth through increased savings and the promotion of capital market development. 17. The Bank's strategy to implement multi-pillar pension reforms was intended to apply to countries that satisfied certain preconditions, including: ( 1)sustainable macroeconomic policies, (2) a sound financial sector, and (3) sufficient implementation capacity. This evaluation uses a set of indicators to assess whether the necessary conditions were in place before Bank support of multi-pillar pension reform. case studies were more likely to rate the Bank's performance satisfactory i n multi- pillar reform countries w h e n these preconditions were met. 18. The Bank only supported PAYG improvements in some coun- tries that did not meet these preconditions, rather than advancing multi-pillar reforms. Insome cases, however, the Bank supported multi-pillar systems incountries with high fiscal deficits, weak finan- cial systems, and ineffective implementation capacity. 19. Although pension shortfalls undermine fiscal stability, the transition costs of immediately switching f r o m a P A Y G system to a f u n d e d system will temporarily increase the fiscal deficit further as the government m u s t continue to p a y pension benefits w h i l e some contributions are diverted into private funds. Thus, countries should first b e advised to achieve fiscal sustainability through expenditure rationalization and revenue reform, including parametric reforms to their pension systems, before embarking on a multi-pillar reform. xviii SUMMARY 20. Despite Strategy's recommendation, the Bank's pension r e f o r m agenda i nclient countries often did n o t include policies to protect the vulnerable elderly who are ineligible for public pensions. F r o m a pov- erty reduction standpoint, countries with l o w pension coverage rates need to explore options t o expand participation i n the formal system and/or p r o v i d e complementary social safety nets t o improve the wel- fare o f the aged. For example, i n LAC, where coverage i s low, pen- sion r e f o r m will assist far fewer future retirees than in ECA, where a high proportion of workers are covered by formal plans. What Has Been the Impact of Reforms? 21. Pension reform requires m a n y years of implementation before its impact can b e fully evaluated. M o s t reforms are too recent for a longer t e r m assessment. Consequently, it is necessary to use indirect indicators to gain insight into the progress t o w a r d achieving the Bank's objectives for pension reform. 22. Large fiscal deficits, stemming i np a r t from imbalances in pen- sion revenues and expenditures, motivated many countries to seek Bank assistance in reforming their pension systems. While the Bank's reforms i m p r o v e d the financial balance of many PAYG systems, addi- tional reforms often were needed, but not enacted, to ensure full fiscal sustainability . 23. One advantage of a multi-pillar system i s to reduce financial risk to future pensioners through portfolio diversification, i n c l u d i n g the existence of public and private components. Investments i n pri- vately funded pillars. Inm a n y countries with multi-pillar systems, however, are not w e l l diversified, although rates of r e t u r n are high as a result of investments i n government bonds. While these bonds offer high returns, they often just compensate for macroeconomic a n d in- vestment risk. I naddition, privately funded systems remained open t o political influence, just l i k e PAYG plans, particularly intimes of economic crisis. 24. Empirical evidence suggests that the secondary objectives of privately funded pension plans to increase savings, develop capital markets, a n d increase worker participation in the pension system have remained largely unmet. There i s little evidence that privately funded pillars have succeeded i nincreasing national savings or de- veloping capital markets. Furthermore, multi-pillar reforms have not increased pension coverage i n most reforming countries. xix SUMMARY Has Adequate Attention Been Given to Institutional Capacity? 25. World Bank operations have supported countries buildinginsti- tutional capacity throughout the pension reform process. Out of more than 200 loans and credits, 129 have supported institutional capacity, including improving the administration of unfunded systems, actuarial forecasting, and regulation of privately funded plans. 26. The need to develop effective pension administration, h o w - ever, has been greater than the assistance provided. The Bank under- estimated institutional weaknesses because of incomplete needs as- sessments, reluctance o n the p a r t of some agencies t o open a dialogue with the Bank, a n d insufficient Bank expertise o n the administration of publicly managed, u n f u n d e d plans. Inaddition, administrative projects that were undertaken would have benefited from better Bank a n d client supervision, particularly i n countries with capacity con- straints. 27. Bank loans to establish regulatory systems for privately nnumber and scope. Inparticu- f u n d e d pensions have been l i m i t e d i lar, i n LAC, investment restrictions m a y have created an additional investment risk for future retirees. But i n some cases successful ad- ministrative a n d regulatory reforms were supported by policy dia- logue rather than investment projects or technical assistance. This was true i n some ECA and LAC countries. 28. The Bank developed a long-term forecasting m o d e l for pen- sions, the Pension Reform Options Simulation Toolkit (PROST), as an in-house tool for policy analysis to help client countries develop fi- nancially sustainable pension systems. However, the Bank’s technical assistance did not develop sufficient local expertise to assess the fiscal balance o f pension programs o n a n ongoing basis or update policy re- forms. Some countries c o u l d not implement PROST because of too f e w trained professionals. Others found PROST data requirements prescriptive or too inflexible to use for country-specific applications. Has Support Been Well Coordinated? 29. The World Bank’s internal and external relationships have af- fected the outcomes of Bank activities and the success of the reforms, through: ( 1)coordination among Bank units and teams, (2) coordina- tion with other donors and international organizations, and (3) rela- tionships with clients. The Bank has yet to develop a decision-making process that is well-coordinated across the three primary Networks in- volved inpension reform (HumanDevelopment, Poverty Reduction and Economic Management, and Finance). Without consistent guide- lines and benchmarks for pension activities, staff changes within the Bank and i nclient countries l e d to inconsistent advice and support over xx SUMMARY time. Furthermore, the Bank has not always been steadfast or efficient nits provision of funding, resulting i i nover-funding or under-funding of particular pension reforms. 30. I m p r o v e d cooperation with international agencies a n d bilat- eral donors over the years has resulted i n stronger pension reforms. But the Bank could still benefit from finding further common ground with i t s international partners despite differences i n perspective. The Bank’s own stop-and-go tactics- that is, lack of sustained within- country attention over several CASs - tended to reinforce discontinui- ties i n the progress of reform. While the Bank w o r k e d successfully with many governments, i t needs to ensure that i t involves a l l rele- v a n t ministries and stakeholders. Summary and Conclusions 31. The Bank’s multi-pillar strategy is w e l l documented with a strong legacy of operational work, economic and sector work, training, and seminars. Reforms have differed regionally and by country, as a result of client concerns and Bank experience. Nonetheless, the Bank’s advice has n o t always been effective. While formal pension systems i n many countries contributed to ballooning budget deficits, the Bank’s preoccupation with fiscal sustainability tended to obscure the broader goal of pension policy, that is, to reduce poverty and improve retire- m e n t income adequacy within a fiscal constraint. 32. To improve this process, recommends: DEVELOP TO DESIGN GUIDELINES PENSION REFORMS AND PAY GREATER TO PARAMETRIC REFORMS ATTENTION a. Pay greater attention to parametric reforms to ensure fiscal sustainabil- ity and to the macroeconomic, financial, and institutional sector pre- conditions necessay for a multi-pillar reform. This would involve pre- paring and implementing guidelines to ensure well-tailored assistance to country conditions a n d consistent policy prescrip- tions including statistical indicators and in-depth assessments. b. Be more realistic i n presenting the benefits of the secondary objectives of pension reform in dialogue with client countries, as there i s insuffi- cient empirical evidence to support the claims that funded systems have or can improve savings and capital market development. BUILD CLIENTCAPACITY c. Develop a checklist for client capacity requirements (including contri- bution collection, contributor database development, actuarial and policy analysis, regulation of multi-pillar operations) t o assess cli- xxi SUMMARY ent requirements and determine how best they can b e met. This would involve ensuring that a plan for technical assistance i s put i n place for reform initiatives so that client capacity i s developed. CONDUCT RESEARCH ON OUTSTANDING k i U E S d. Ensure that adequate analysis i s conducted on key issues such as in- come of the aged, the impact o f corruption and governance on the feasibility of effective pension regulation, methods to foster com- petition among pension funds, guidelines for investment alloca- tion, the design of non-contributory systems, and ways inw h i c h capital markets develop, as w e l l as research offering cross-country evidence o n these topics. IMPROVE INTERNAL AND EXTERNAL COORDINATION e. Develop a process to ensure that cross-sector issues are considered in- c l u d i n g financial issues identified by the FSAP and maintain closer coordination between the Development Economics vice presidency, the Networks, sector units, and country units. f. Develop a strategy to play a greater role in consensus building among stakeholders, in particular, other international organiza- tions and client agencies. xxii 1. The Strategy for Pension Reform 1. 1 Pension reform i s a focus of World Bank activities because pen- sions are an important p a r t of the social safety net for workers covered by the formal pension system i nmany client countries, providinga mechanism to reduce the risks of old-age poverty a n d a means to smooth lifetime income to maintain living standards in retirement. Pensions are only one p a r t of the sdety net to protect the aged, w h i c h may include other public programs, such as targeted benefits for the aged poor or universal benefits for a l l the aged, and a host of informal arrangements, including direct family support. Pension systems must be fiscally and politically sustainable to achieve their income-support objective. Unsustainable pension systems can b e a n obstacle to fiscal stability, economic growth, and poverty reduction. 1.2 The need for pension r e f o r m has become pressing over the past decades as demographic aging has strained pension systems around the world, leading to large expenditures, large deficits, and high contribution rates. Even countries with relatively young popula- tions have experienced these problems due to high benefits and lax eligibility requirements. In addition, poor administrative capacity a n d practices have resulted i n ineffective collection, entitlement, a n d bene- fit determination. Inmany cases the pension system has become a source of fiscal and macroeconomic instability, a constraint to eco- nomic growth, a n d a n ineffective and/or inequitable source of retire- m e n t income. E v e n c i v i l service pensions in countries with no other pension system m a y become a fiscal drain on government resources. 1.3 The World Bank has been a leader i nassisting countries in The Bank has pension reform. Since 1984, the Bank has helped 68 countries r e f o r m been a leader in naddi- their pension systems with more t h a n 200 loans and credits. I pension reform, tion, the Bank has issued over 350 papers and publications on pension offering both reform. This report i s the first comprehensive evaluation of the Bank’s lending and involvement, assessing the relevance of the Bank’s strategy and the advisory resulting development outcomes. services 1.4 During the 1990s, the Bank was criticized for following a dogmatic approach, providinglittle support for the improvement of public systems and aggressively promoting the privatization of social security, regardless of the country’s characteristics a n d i n i t i a l condi- tions. Critics claimed that the Bank oversold the benefits of multi- pillar systems, particularly the benefits of a n e w second pillar, w h i l e 1 CHAPTER1 THESTMTEGY FOR PENSION REFORM The Bank has simultaneously underestimating the advantages of p u b l i c l y managed been criticized programs. Supporters of the Bank’s approach stress the balance of as- for aggressively sistance w h i c h covered both single- and multi-pillar reforms a n d sug- promoting the gest that pension r e f o r m failures have been primarily the result of in- privatization of adequate government policy. social security 1.5 The 2001 publication Social Protection Sector Strategy: From Safety Net to Springboard (Strategy)i s the Bank’s official strategy on pension reform, supporting a multi-pillar framework as best practice if proper initial conditions are in place. Strategy f o l l o w e d the Bank’s 1994 policy research report Averting the Old Age Crisis (Averting), w h i c h set an agenda for pension reform and p r o v i d e d the intellectual underpinnings to m u c h of the Bank’s activities i n the 1990s. Averting proposed a similar, but m o r e detailed strategy; i t has been influential w o r l d w i d e as a blueprint for pension r e f o r m a n d i s w i d e l y perceived as representing the Bank‘s thinking, especially throughout the 1990s. Nevertheless, Averting was never presented or agreed upon as a sec- tor strategy with the Bank’s Board 1 (see Box 1). 30X 1. Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth Strategy praised Averting the OZd Age Crisis for being the first comprehensive diagnosis of pension programs for recommending greater reliance o n private-sector investment man- 3gement to address the challenge o f ( 1)demographic trends that undermine fiscal sustain- 3bility, and (2) govenunent policies that are subject to political pressure. Averting argued that the best way for most countries to meet the challenges of an aging w o r l d is through a multi-pillar system witlx 0 A mandatory tax-financed public program designed to alleviate poverty; 0 A mandatory funded, privately managed program (based o n personal savings ac-’ counts or occupational plans) for savings; a n d 0 A supplementary voluntary option (through personal saving or occupational plans) for people w h o w a n t more protection. Averting proposed four alternatives for the first two pillars: 0 A mandatory personal saving p l a n with a flat benefit public scheme; 0 A mandatory personal saving plan with a minimum pension guarantee i nthe pub- lic scheme; 0 A mandatory occupational plan with a flat benefit public scheme; a n d 0 A mandatory occupational p l a n with a means-tested public scheme. Averting discourages the use o f an earnings-related scheme for the public pillar, but i f one is provided, ”the wage replacement rate should be based on lifetime earnings.. .[italics i n original].” Averting notes that the ”right mix” of pillars is not the same at all times and places, but depends o n a country‘s objectives, history, and current circumstances. Avert- ing suggests adopting a slow reform process for formal systems i nlow-income countries, and that public programs i n the r u r a l areas of poor countries ”should concentrate o n so- cial assistance for the neediest of a l l ages, while every effort is made to develop the ca- pacities that will enable more complex formal systems to w o r k well. Mandatory contri- bution programs should be introduced first in the formal labor markets of urban areas, where the informal system is most likely to have broken down.” 2 CHAPTER1 THESTRATEGY FOR PENSION REFORM The Social Protection Sector Strategy 1.6 Strategy outlines a social r i s k management framework spelling Bank strategy out h o w public safety n e t programs can cope with, mitigate, or pre- supports a vent the risks that increase a population’s vulnerability to poverty.2 flexible, multi- This document i s the basis for the Independent Evaluation Group () pillar framework pension evaluation, as it provides the official strategy for Bank opera- tions. It supports flexible multi-pillar pension r e f o r m w h i l e ensuring adequate retirement income for i n f o r m a l sector workers a n d the life- time poor. 1.7 Strategy offers a clear description of the purpose and function of pension systems, indicating that the improvement o f old-age in- come security inthe f o r m a l sector begins with ”a flexible ap- proach.. .focusing on a ’multipillar’ system that many countries throughout the world are successfully implementing.” Further, w h i l e ”maintaining t h i s approach, the m a i n challenges are to ensure ade- quate retirement income for i n f o r m a l sector workers a n d lifetime poor people, as w e l l as for particularly vulnerable groups such as widows, by strengthening their access to earnings, savings a n d other assets.” This is outlined m o r e succinctly in Figure 1 , w h i c h i s the framework used for this evaluation. 1.8 As explained i n Strategy, a multi-pillar system consists of three The framework types of income support: (1)a publicly managed, unfunded, defined has a public benefit (DB) program; (2) a privately managed, fully funded, defined unfunded pil- contribution (DC) plan; and (3) voluntary retirement savings in the lar, a private form of housing, insurance or other assets. The first pillar addresses funded pillar, poverty alleviation and the second provides consumption smoothing. and a volun- Incontrast to a single public program, two pillars are expected to tary pillar safeguard against the costs o f an aging population, protect the system n the from political risk, a n d facilitate i n d i v i d u a l decision-making i process. Inaddition, multi-pillar systems are expected to contribute to national savings and financial-market development. 1.9 Strategy indicates that w h i l e the proposed multi-pillar ap- Bank strategy proach to pensions ”continues to b e a useful benchmark, i t is not a suggests using blueprint [italics added], a n d any r e f o r m has to take account of a coun- the multi-pillar try’s starting conditions a n d preferences.’’ Strategy promises that the approach as a ”World Bank’s future w o r k on pension r e f o r m will focus m o r e on the benchmark, not provision of retirement benefits for people i nthe i n f o r m a l sector and a blueprint o n old-age income support for the life-time p o o r through public non- contributory schemes a n d community support.” 1.10 Strategy specifies that certain conditions m u s t b e f u l f i l l e d for the Bank to support a multi-pillar reform, including objectives for in- come redistribution, macroeconomic feasibility, financial sector readi- ness, a sound regulatory and supervisory framework, a n d sufficient 3 CHAPTER1 THESTRATEGY FOR PENSION REFORM Specific social, administrative capacity. Criteria to judge the soundness of a pension macroeconomic, reform include: (1)distributive effects on the income of the aged, and financial (2) the macro and fiscal policy environment, (3) the capacity of the conditions must administrative structure to operate a multi-pillar system, and (4) the be met before soundness o f regulatory a n d supervisory arrangements. implementing multi-pillar reform Secondary Obiectives Primary Obiectiies {economic qrowth) 1. Poverty reduction among Goals 1. Increased national savings - b the aged 2. Capital market growth 2. Lifetime consumption 3. Labor market incentives -_ smoothing Development A sustainable Outcomes multi-pillar system in informal sector Desinn Muiti-pillar System 1. Publicly managed, unfunded defined benetit scheme Improve PAYG 1 outputs 2. Privately managed, funded system defined contributin scheme 3. Voluntary retirement savings Reform Initial Conditions 1. Macroeconomic environment Activities 2. Financial sector 3. Governance Bank Products 1. Analytical and advisory analysis Inputs 2. Loans and credits 1. 1 1 Because Averting was not a formal Bank strategy, Bank staff h a d considerable leeway throughout the 1990s i n developing country-based solutions. While Strategy does not indicate w h i c h financial sector pre- conditions are required for a successful multi-pillar reform, related 4 CHAPTER1 THESTRATEGY FOR PENSION REFORM World Bank analysis (Vittas 1998) i s more specific.3 Strategy also does not provide the criteria against which to evaluate multi-pillar reform success, including improvements inretirement income security, gains infiscal stability, and increases in savings and capital market develop- ment. 1.12 The relevance of the Bank’s strategy. The Bank‘s strategy for Bank strategy pension reform i s highly relevant, as formal pension systems are a n on pension important means to reduce poverty among the aged and ensure life- reform is time consumption smoothing inmany W o r l d Bank client countries. relevant to When pension expenditures exceed revenues, the difference will in- poverty crease the consolidated government deficit, a l l else remaining un- reduction changed, threatening macroeconomic stability and undermining re- among the aged tirement income security.4 For those outside the formal pension system, the Bank’s strategy recommends designing policy to either in- crease coverage or offer non-contributory schemes to allow better risk management for uncovered workers. 1.13 Formal public pension programs are more important in some regions than in others. For t h i s reason, the recognition of the need to reduce poverty among the non-pensioned aged is a n important com- ponent of the Bank’s strategy. For example, pension coverage i s great- est inthe Europe and Central Asia Region (ECA) at a n estimated 60 percent of the working-age population; compared to L a t i n America a n d Caribbean (LAC) a n d M i d d l e East a n d N o r t h Africa (MNA) where i t i s closer to 30 percent, East Asia and Pacific (EAP) where i t i s a little over 15 percent, and sub-Saharan Africa (AFR) and South Asia (SAR) where it i s less than 10 percent.5 1.14 Countries that have an increasingly high percentage of the Demographic population reaching retirement age m a y face severe future fiscal imbal- aging may lead ances. These countries are also more likely t o have high pension cover- to severe fiscal age rates, and as a result, the Bank’s strategy o n pension reform will imbalances in likely affect a large portion of the population. ECA countries are a the future prime example. Even countries with lower coverage and younger populations, including countries in LAC and other Regions, face fiscal issues sirnilar to countries with serious demographic aging problems, particularly w h e n employment i n the covered sector i s declining rela- tive to an increasing number of retirees. I nthose countries, the rele- vance of Bank’s strategy is also evident. Inregions such as Africa, pen- sion reform has been less of a Bank priority. 1.15 Pension r e f o r m continues to b e a topic of some contention Bank thinking among researchers inside a n d outside of the Bank f r o m a theoretical on pensions has a n d practical perspective (see Appendix A). Since 1994, W o r l d Bank continued to thinking o n pensions has continued t o evolve as pension issues are evolve as debated, and experience o n the topic has become more expansive. pension issues This i s evidenced in the edited volume New Ideas about O l d Age Secu- are debated 5 CHAPTER1 THESTRATEGY FOR PENSION REFORM rity (2001), w h i c h contains an evaluation highly critical of Averting, a n d the 2005 report Keeping the Promise o f SocialSecurity in Latin Amer- ica (Promise), an assessment of pension r e f o r m i nLAC with recom- mendations for a regional strategy. Pension policy also received atten- n the 2004 report Economic Growth i n the 1990s: Learningfiom a tion i Decade ofReform (Learning). M o s t recently, in 2005 the Bank released a major research report devoted t o pensions, OZd Age Income Support in the 21st Centuy: The World Bank's Perspective on Pension Systems and Reform (Perspective).6 The topics, emphasis, a n d findings of these re- ports differ, indicating the nature of the debate within the Bank, a n d reflecting a diversity of conclusions about the outcome of Bank- supported pension-reform activities over time. The Structure of the Report The report 1.16 Subsequent chapters use statistical indicators, assessments evaluates Bank from reviews of Implementation Completion Reports (ICRs) and Pro- strategy, ject Performance Assessment Reports (PPARs), assessments from the whether it was Financial Sector Advisory Program (FSAP), interviews with Bank staff followed, and a n d external stakeholders, desk reviews, and 16 case studies to evalu- how well Bank ate the quality a t entry and development outcome of the World operations were Bank's pension strategy.7 implemented 1.17 Chapter 2 assesses whether the Bank f o l l o w e d i t s strategy by reviewing i t s lending operations and non-lending activities, i n c l u d i n g economic and sector work (ESW), policy dialogue, and training and dissemination. The chapter also examines the outcome ratings for pro- jects with pension components. Chapter 3 assesses whether Bank lending decisions were based on best practice guidelines. Chapter 4 analyzes whether Bank-assisted reforms achieved their p r i m a r y objec- tive of providinga fiscally sustainable pension system a n d their sec- ondary objectives of increasing savings and developing capital mar- kets. Chapter 5 examines the Bank's assistance i nbuilding capacity for administration, regulation, supervision, and actuarial analyses. Chapter 6 evaluates the Bank's internal and external coordination, as w e l l as the influence of exogenous factors on project outcomes. Chap- ter 7 summarizes the findings and presents specific recommendations, including the establishment of f o r m a l guidelines to create a n objective a n d coordinated method to evaluate Bank strategies for pension re- form. 6 CHAPTER1 THESTRATEGY FOR PENSION REFORM 1.Some of the differences between Strategy and A v e r t i n g arose as a result of the ex- NOTES periences the Bank gained through its involvement with countries that underwent multi-pillar and parametric reforms. 2. Strategy provides a considerable discussion of the types of risks and risk manage- ment. Pensions are risk-coping (poverty alleviation) and risk-mitigation (a drop in income after retirement) mechanisms (World Bank, Social Protection Sector Strategy, 2001). n Perspective (2005). For additional discussion, see Chapter 3. 3. This work was cited i 4. The sustainability of a pension system is at risk when the proportion of active work- ers to retirees is low, particularly in PAYG contribution-based systems. This is easy to demonstrate based o n a simple formula, as the revenues from contributions must equal the wage bill times the contribution rate, while pension expenditures equal the number of pensioners times the average pension. W h e n pensions are high relative to wages, the contribution rate will need to be higher. When there are fewer workers for each retiree, the contribution rate will also need to be higher. While countries may also have accu- mulated additional assets to cushion the shortfall, i f demographic factors are important, this is only a temporary solution to fiscal inkolvency. If there is a limit o n affordable contribution rates, benefits will have to be reduced. 5. See Appendix E for projected coverage rates for a set of World Bank client coun- tries. The projections are based o n an equation regressing actual coverage rates (Palacios and Pallares-Mirrales 2000) o n GDP per capita interms of PPP, the percent- age of t h e population over age 65, and regional dummies. The adjusted R-squared statistic was 0.90 indicating a statistically close fit. 6. This volume explains the Bank's perspective o n pension reforms but, as also noted in the Preface, it has not undergone the review accorded to official World Bank publi- cations. 7. The case studies include six countries inthe L A C Region, eight countries in ECA, and two countries inEAP. The selection of countries was determined by the Bank's in- volvement inpension reform in each country, the regional implementation of multi- pillar systems, exclusively inECA and LAC, and the fact that most pension activities have taken place inECA and LAC. Specific countries selected for in-depth appraisals of the Bank's assistance to pension reform comprised a substantial share of the value of the pension portfolio in each region (more than 75 percent of the value of the pension com- ponent) and constituted a representative sample (the selection criteria included the subregion, size of the population, level of per capita income, level of financial sector de- velopment, and other factors). Activities in other regions are assessed through desk re- views and interviews with Bank staff. 7 2. The Bank’s Support for Pension Reform 2.1 The World Bank supported pension reforms in 68 countries through l e n d i n g and analytical and advisory analysis (AAA) - including ESW (see Box 2), policy dialogue, seminars, and training. This chapter first reviews the Bank’s AAA to assess i t s thoroughness in elaborating the Bank’s pension strategy. It describes reforms the Bank supported or advocated to determine whether the Bank fol- l o w e d i t s own flexible multi-pillar model. Finally, the chapter reviews Bank-supported projects to determine whether they h a d satisfactory development outcomes. 1 Box 2. Identifying Pension-RelatedEconomic and Sector Work In addition t o the W o r l d Bank lending projects, the pension database con- tains 355 ESW studies with substantial analyses o n pensions and the income of the aged. ESW was selected based o n the relevance and depth of analyses o n pensions f r o m nearly 1,000 documents in ImageBank containing the key words ”pension,” “social security,” “contractual saving,” and ”provident funds” as of July 2004. Although this literature spans fiscal 1962 t o fiscal 2004, this report focuses primarily o n reports after fiscal 1990. The pension database also identifies whether each ESW study addresses the commonly mentioned pension-related topics. The topics are: (a) poverty, (b) gender impact, (c) income of the aged, (d) targeting/coverage/eligibility, (e) fiscal sustainability, (f) capital market development, (8) contractual savings, (h)fund management/ investment, (i) annuities/ insurance, (j)pension sys- tem description, (k) pension reform design, (1) public information/ political support, (m) private fund, (n) public administration/costs, ( 0 ) transition costs. 2.2 The Bank‘s extensive ESW addressed a broad range of pension issues. Analysis in several areas, however, especially income of the aged a n d financial sector development, lacked sufficient depth to as- sist in project preparation. Conferences and seminars, w h i c h initially promoted Averting, broadened in content to include a fuller range o f pension topics, congruent with changes in ESW. Staff interviews sug- gest the impact of the Bank’s i n f o r m a l policy dialogue on pension re- form also has been influential. 2.3 Backed by a compendium of ESW and training, Bank loans and credits have supported multi-pillar reforms in ECA and LAC, which, 9 2 CHAPTER SUPPORT FOR PENSION REFORM THEBANK'S consistent with Strategy, differ considerably by country. The reason for the variation, however, could b e due to specific country preferences, specific country conditions or exogenous factors., While the majority of the development outcome ratings for the pension components of Bank loans are satisfactory, satisfactory outcomes for i n d i v i d u a l projects do nsatisfactory development outcomes for the n o t necessarily result i Bank's pension reform work overall. Analytical and Advisory Services Bank analytical 2.4 The Bank's AAA includes published economic a n d sector and advisory work, informal and formal policy dialogue, and World Bank training services have a n d seminars.' All three forms o f AAA influence the direction of Bank helped operations by p r o v i d i n g information on pension r e f o r m issues to determine the Bank staff, Bank client countries, and the wider public, i n c l u d i n g direction of other stakeholders and donors. assistance 2.5 Economic and Sector Work.2 The World Bank's papers a n d pub- lications on pensions constitute a n extensive technical foundation adding to a n d deepening the understanding of pension r e f o r m i n more than 66 countries. Over 200 Bank country studies comprise a comprehensive analytic base to examine the goals, preconditions, a n d underlying principles of the Bank's evolving pension strategy re- form.3 Pension studies have taken a variety of forms i n c l u d i n g papers from m a n y of the World Bank sector units, most prominently the Pen- sion Primer Series produced by the Social Protection Network. 2.6 On a regional basis, the preponderance of studies focus on Bank ESW ECA, followed by LAC, the only two Regions that have undertaken focuses mainly multi-pillar pension reforms. Over 40 percent of the studies address on countries pension issues in countries that adopted multi-pillar reforms, and a that have dozen countries account for over four-fifths of the studies. M o r e than adopted multi- 10 studies each have been issued for Argentina, Brazil, Poland, and pillar reforms in Russia. Countries with f u n d e d pillars average over four studies ECA and LAC apiece, and countries without f u n d e d pillars average just over two studies apiece. Brazil, w h i c h has not implemented a f u n d e d pillar, i s a n exception, with 16 studies. Analyses have focused on 2.7 World Bank ESW includes analysis of specific pension topics, fiscal issues descriptions of pension systems in developed economies, cross-regional because fiscal analysis, and country assessments. Six topics are the most prevalent: imbalance often (1)fiscal sustainability and transition costs, (2) regulation of f u n d e d triggers pension systems, (3) pension coverage, (4) living conditions of the aged, reform (5) capital markets, and (6) the administration of public systems (Figure 2).4 Fiscal issues are the focus of substantial analysis because fiscal im- balances are a primary reason for the Bank to assist inpension reform. Inmany instances, this has been the chief reason to shift to a multi- 10 CHAPTER2 THEBANK’S SUPPORT FOR PENSION REFORM pillar system. Bank ESW has also been considerable i nassessing the regulation of private pensions, including asset allocation, contribution collection, and governance. Figure 2. More Sector Work Addressed Fiscal Issues and Transition Costs and Fewer Reports Discussed Public Administration Fiscal sustainabilityltransifon costs RegulaSon of funded systems Pension coverage Living conditions Capital markets Administrafon of public systems , , , , I 0 10 20 30 40 50 60 70 80 90 Percentage of countries Source: analysis of Internal Bank data. 2.8 Although the Bank’s pension library i s substantial, ESW often lacks the detailed analysis needed to assist i nproject implementation. For example, although many countries have h a d poverty assessments, the analysis of the income of the aged i s limited.5 Poverty assessments do n o t relate the risk of poverty among the elderly t o age, family structure, gender, or location- a l l necessary to understand the role n establishing a safety net.6 Pockets o f poverty or pensions m a y p l a y i sources o f income are not considered; distributional data are not pro- vided. The impact of gender on the welfare of the elderly i s assessed in only 1 1percent of countries with pension ESW. 2.9 Similarly, although pension coverage is a n important topic, lit- Although ESW tle empirical research has been conducted on the limits o f formal pen- covers many sion coverage or ways to increase it, despite interest expressed in Stvat- topics, some egy to expand upon this topic.7 Furthermore, ESW has been l i m i t e d on analyses raise topics closely related to o l d age pension reform, including disability more issues and survivor’s pensions, public information, and the political process. than solutions Inaddition, there is a need for more research on pay-as-you-go (PAYG) administration, including collection and payment strategies, despite its importance to both PAYG systems and multi-pillar reforms.8 2.10 Country financial market conditions, a key determinant of ncoun- readiness for multi-pillar reform, have received little attention i try reports, although the Bank has published seminal findings on the issue. The majority of country-specific investment-related ESW encour- 11 CHAPTER 2 SUPPORT FOR PENSION REFORM THEBANK'S ages the use of pension funds to improve capital markets and/or sug- gests that contractual savings mechanisms should b e developed, with- out assessing financial-market stability.9 Thus, the Bank's ESW implic- itly assumes that capital market development will follow pension reform; that is, the supply of funds will create its own demand. Less at- tention has been focused on financial and capital market development inECA t h a n elsewhere, including MNA and LAC, Dissemination 2.11 Training and Dissemination.The influence of seminars and workshops, particularly those conducted on a w o r l d w i d e or regional of Bank ESW initially focused basis, has h a d a substantial impact on policymakers, although this entirely on impact i s difficult to document. The dissemination of Bank's research Averting the Old in the early years focused on Averting. Between 1994 a n d 1999, the Age Crisis but Bank sponsored a series of promotional tours i n c l u d i n g more t h a n 100 was extended seminars a n d presentations o n Averting a n d related research. It i s no thereafter surprise, therefore that most o f the world considers Averting t o be the Bank's pension model. Subsequently, the agenda of seminars and workshops broadened, with the number of course offerings expand- ing from one i n1996 to 13 i n2003, reflecting the diversity of reforms taking place in ECA and LAC. World Bank Institute (WBI) seminars a n d workshops include basic education on multi-pillar reforms, train- ing o n the Bank's pension simulation m o d e l (PROST), a n d training on special pension reform topics such as administration and regulation. Recently, the WBI has shifted from global programs t o w a r d m o r e country-focused training on regional problems. Policy dialogue 2.12 Policy Dialogue. Policy dialogue has been particularly important for the Bank's w o r k o n pension reform, but written documentation i s generally supported multi- limited, since ina number of instances the dialogue was informal or did nloans or credits. Nonetheless, its influence should not b e n o t result i pillar reforms underestimated. While Bank dialogue generally supported the multi- pillar strategy, it has n o t been consistent across countries or within countries over time i nits consideration of the preconditions for multi- pillar reform, especially i nECA.10 A l t h o u g h the Bank strongly sup- ported multi-pillar reform i nLAC, it has h a d a stop-and-go dialogue i n a number of countries, w h e n economic circumstances were essentially unchanged?' InAfrica, Bank discussions with clients on the appropri- ateness of multi-pillar systems have been inconsistent across countries with similar macroeconomic, social, and financial conditions.l2 Lending Operations in Support of Pension Reform 2.13 W h i l e the Bank has not used a one-size-fits-all approach to pension reform, it has concentrated on multi-pillar systems rather than PAYG alternatives or non-contributory schemes. 13r14 Little sup- port was p r o v i d e d to expanding old-age benefits to workers i n the in- f o r m a l economy. Strategy emphasized the importance of this t y p e of 12 CHAPTER2 SUPPORT FOR PENSION REFORM THEBANK'S intervention. InECA, where countries were more likely t o already have h a d old-age social assistance, reforms were more likely t o b e ho- listic, that is, i n c l u d i n g a full assessment of other social protection programs s u c h as social assistance for the aged. In AFR, the Bank p r o v i d e d s m a l l loans to a number o f countries. Except for two large loans t o Korea, Regions other than ECA and LAC received only a few small loans for pension reform. 2.14 Over the past two decades, the W o r l d Bank provided over 200 The Bank has loans and credits with components supporting pension r e f o r m to 68 helped 68 countries.'5 T o t a l lending amounted to $34 billion. ECA and LAC countries dominated, with 40 countries in these two Regions receiving roughly through more $11 b i l l i o n each. Bank funding was most active during fiscal 1998- than 200 loans 2001, w h e n operations with pension components totaled $19 b i l l i o n or and credits 56 percent of a l l operational pension spending (Table 1). Table 1. ECA and LAC Received More Support for Pension Reform Than Other Regions from Fiscal 1984 to 2005 Size of pension No. of No. of Commitments component Region countries projects ($ bn) ($ bn) AFR 14 26 $1.5 $0.1 EAP 4 7 $7.4 $0.5 ECA 25 93 $10.8 $1.5 LAC 15 57 $10.7 $3.1 MNA 6 9 $1.I $0.1 SAR 4 12 $2.7 $0.1 Total 68 204 $34.2 $5.4 Source: analysis of internal Bank data. 2.15 A s support for pension r e f o r m generally constituted only a p o r t i o n of each loan or credit operation, calculated the pension- specific component for each project (Box 3). 16 Pension-specific lending undertaken by the W o r l d Bank totaled $5.4 billion. ECA a n d LAC dominated commitments for pension reform, but LAC accounted for a higher pension-component share a t about 40 percent o f the total l o a n or credit. 2.16 W o r l d Bank pension projects include specific policy reforms More than four. for b o t h PAYG a n d multi-pillar systems. A l t h o u g h more t h a n four- fifths of all Bank fifths of a l l Bank loans supported PAYG reforms, nearly one-third of loans supported those also supported funded second pillars as p a r t o f a multi-pillar re- pay.as.you.go form, and nearly one-third supported voluntary pensions. Overall reforms more than three-quarters o f a l l projects related to multi-pillar pension reform also included a PAYG component. Bank l e n d i n g for second pillar reforms was provided without support f o r first p i l l a r assistance 13 CHAPTER2 THEBANK’S SUPPORT FOR PENSION REFORM in only three countries; none of these satisfied the precondition for multi-pillar reformJ7 Countries implementing multi-pillar systems re- ceived h a l f again as m a n y loans for PAYG reforms as countries stick- ing with their P A Y G systems, a n d over twice as many resources for the PAYG pension component. Box 3. How Identified Pension Projects screened project documents for 400 loans and credits and identified more than 200 projects containing a pension component for fiscal years 1984 to 2005, developing a reliable and comprehensive database of W o r l d Bank pen- sion projects. For adjustment projects, the pension component i s identified as the relevant condition i nthe policy matrix; for investment loans (including technical assistance), the component described in the detailed project de- scription. also compared its database with existing regional pension data- bases t o ensure completeness. The classification of projects was n o t always straightforward. The pension component consists o f general analytic sup- port, actual reform measures, and institutional capacity building. took a more comprehensive approach including Bank projects that specified a clear intent t o reform a country’s pensions through exploratory measures or re- search (classified as general analytic support). M o s t of this report focuses o n projects containing specific measures to reform legislation (actual reform measures), however, and specific types of technical assistance (institution capacity building).Inmost cases, the loans and credits were used to reform pension design, although the database also includes loans t o p a y off pension arrears in certain parastatal enterprises (e.g., coal, railroad, etc.). 2.17 Countries legislating a n d implementing multi-pillar systems also received m o r e loans per country t h an others (Table 2). Nearly Countries with three-quarters of pension loans w e n t to countries in ECA a n d LAC, multi-pillar the o n l y Regions enacting multi-pillar reforms. Further, of the 23 systems countries receiving four or more pension loans, 13 enacted second pil- received more lars. Only one country, Georgia, receiving World Bank assistance for loans than second pillar r e f o r m failed to pass the necessary legislation to imple- others m e n t the pillar. Second pillar 2.18 Second-pillar assistance was concentrated and substantial. Of assistance has 1 1countries receiving funding of m o r e than $100 million apiece)* been eight enacted mandatory funded pension laws.l9 Overall, countries concentrated with second pillar legislation accounted for over h a l f of the $5.5 bil- and substantial lion funding for pensions projects. M e d i a n World Bank l e n d i n g per country implementing second pillar reforms w a s $50 million, com- pared to $7 million for those not implementing second pillars.20 14 2 CHAPTER THEBANK’SSUPPORT FOR PENSION REFORM Table 2. Countries with Multi-pillar Systems Received More Assistance Average size of Total Total Size Average pension Pension system commit- of pension no, of component Loan pillar type in recipient No. of No. of ments component. projects per country type country countries projects ($ bn) ($ bn) per country ($ mm) Pillar 1 PAYG system 45 100 13.5 2.4 2.2 53.7 Multi-pillar system 21 70 13.5 2.7 3.3 126.4 Pillar 2 Multi-pillar system 20 43 5.2 1.7 2.2 84.4 Note: Most projects support more than one pillar, so the figures do not add up to 100%. The first pillar is defined as a publicly managed, unfunded, defined benefit pillar. The second pillar is defined as a privately managed, funded, defined contribution pillar (with no redistribution). The third pillar is defined as a voluntary, privately managed pillar. In reality, the delineation among many systems is less clear (e,g,, systems that are partially funded or are managed by a tripartite board). In general, this report classifies this category of more ambiguous systems as second pillar if the funds are invested in individual accounts. 2.19 InLAC, the Bank supported multi-pillar reforms, which, i n Many countries one form or another, were implemented i n most countries i nw h i c h that received the Bank h a d a dialogue.21 A m o n g LAC countries with multi-pillar multi-pillar systems, l e n d i n g was concentrated i n six countries: Argentina, Bo- assistance livia, Colombia, Mexico, Peru, a n d Uruguay.22 Eleven ECA countries established implemented multi-pillar reforms with Bank support. In other Re- multi-pillar gions, i n c l u d i n g AFR and EAP, the World Bank p r o v i d e d small systems amounts of technical assistance for multi-pillar reforms that have not yet been implemented. 2.20 PAYG Reforms. W h i l e Strategy recommends the implementa- If macro- tion of multi-pillar systems, it supports parametric reforms w h e n ini- economic and tial macraeconomic and financial sector conditions are n o t i nplace. financial The role of World Bank assistance for PAYG reforms has been to conditions have achieve fiscal sustainability by raising retirement ages, lengthening not been met, minimum contributory periods, restricting pension eligibility and Bank strategy early retirement options, and, occasionally, increasing contribution recommends rates and/or earnings ceilings. InLAC, although the World Bank PAYG reforms supported a combination of PAYG and multi-pillar reforms, it also supported PAYG reforms infour countries providing participants a choice between a reformed PAYG system and multi-pillar option. Ln ECA, the Bank supported a large number of small loans for paramet- r i c reforms w h e n multi-pillar systems were not a n option. 2.21 Types ofMuZti-pillar Reforms. Multi-pillar pension reforms sup- Bank supported ported by the World Bank varied considerably, p a r t l y d u e to country multi-pillar preference and type of pension or social assistance system previously reforms varied n place.23 InLAC, although the Chilean example h a d a substantial in- i considerably fluence, many reforms did not strictly follow i t s example. InECA, where the Bank also supported multi-pillar systems, innovative de- signs with larger P A Y G pillars and notional defined contribution 15 2 CHAPTER SUPPORT FOR PENSION REFORM THEBANK’S (NDC) formulas, were frequently implemented. Other reforms devel- oped quite slowly and some were never implemented. The rationale for 2.22 The rationale for adopting mandatory funded pensions differed adopting in LAC and ECA. InLAC, the primary objective o f World Bank support mandatory was to improve financing and reduce the political influence on pension funded plan operations by replacing P A Y G plans with funded systems. In pensions ECA, the k e y concerns were fiscal stability and demographic pressures, differed by w h i c h were to b e relieved by reducing the size of PAYG components i n region the future and strengthening the relationship between contributions and benefits to encourage participation and equity. 2.23 The m a i n difference between the actual reforms a n d the Bank’s strategy is that most ECA countries maintained a relatively substantial P A Y G pillar in the reformed system, where pensions w e r e also related to contributions, a design not explicitly considered i n Strategy. ECA pension reforms were likely to b e phased i nby age co- hort a n d only made mandatory for younger workers.24 By contrast, w h i l e LAC reforms were more l i k e l y to b e Chilean i n style, m a n y LAC reformers also continued to support single-pillar PAYG systems for substantial portions o f their populations. 2.24 Ina multi-donor environment, it i s difficult to determine whether the r e f o r m design was the result of the Bank taking the coun- try’s specific considerations i n t o account, the task manager’s prefer- ences, or the country’s desire for a specific reform. In terms of non- Bank influences, NDC reforms in Sweden and I t a l y became the m o d e l for a number of countries i n ECA, and the Chilean r e f o r m influenced policymakers in ECA and LAC. I naddition, other donors also influ- enced pension policy design. InLAC, the Inter-American Develop- m e n t Bank, Chilean consultants, a n d the U.S. Agency for International Development (USAID) were particularly important; i n ECA, the European Union and a host of bilateral donors i n c l u d i n g Sweden, the Netherlands, and Denmark p r o v i d e d support and advice.25 Reviewing the Development Outcome of Pension Components About three quarters of the 2.25 Of the 200-plus loans and credits on pension reform, the per- pension formance outcome of 139 projects was rated for the pension compo- components, as nent a n d the project overall. 26 27 Three quarters of pension-component well as the ratings are satisfactory (Table 3). However, the ratings f o r the entire projects they project tend to b e m o r e favorable, with only 13 percent of the projects were part of, n 77 percent of the loans were rated unsatisfactory. W h i l e the ratings i were rated consistent between the pension component a n d project overall, the satisfactory on pension component was l o w e r in 18 percent of the loans. Thus, not outcome every project rated satisfactory overall has a satisfactory pension component. 16 CHAPTER 2 SUPPORT FOR PENSION REFORM THEBANK'S Table 3. Most Development Outcomes for Pension Components Were Satisfactory Pension component rating I Overall project rating Rating No. % of category No. % of category Satisfactory 101 75 122 87 Unsatisfactorv 33 25 18 13 Total 134 100 140 100 2.26 The success or failure o f any i n d i v i d u a l loan does not predict Social the outcome of the full Bank p r o g r a m of activities supporting pension protection and reforms.% Loans l e d by the social protection and financial sectors finance rated were more l i k e l y t o have satisfactory development outcome ratings best among the for the pension component t h a n those l e d by the economic policy sec- sectors, and tor. 29 This is true for multi-pillar systems and P A Y G systems. By re- LAC rated better gion, a higher proportion of loans made in LAC were rated satisfac- than ECA tory than those in ECA.30 While ratings for projects in countries with PAYG are similar i n LAC a n d ECA, ratings in LAC are m u c h higher than those in ECA for projects in countries with multi-pillar systems. Summary and Conclusions 2.27 The Bank has influenced pension reforms around the world through loans and credits and AAA. W h i l e ESW on pensions covered a broad range of topics, it lacked the depth of analysis needed to assist with retirement income policy and financial sector development dur- ing implementation. Country-specific ESW was likely to b e strong in terms of fiscal analysis and funded-pillar regulation, reflecting the multi-pillar focus of the Bank's pension r e f o r m strategy, but i t did not cover a l l of the necessary topics. 2.28 The strengths a n d weaknesses of Bank lending parallel the strengths and weaknesses of Bank AAA. While satisfactory ratings for most i n d i v i d u a l loans are encouraging, one-fourth of pension compo- nent outcomes were unsatisfactory. Multi-pillar pension reforms supported by World Bank lending varied by country, but this m a y re- flect i n d i v i d u a l country conditions or the influence of exogenous fac- tors. Overall, the Bank offered greater resources to countries develop- ing multi-pillar systems a n d less assistance in assessing or developing old age programs for uncovered workers. NOTES 1.See Appendix G for a listing of publicly available World Bank pension ESW. 2. Written AAA includes core ESW, LAC regional studies and FSAP reports done in conjunction with the IMF. 17 2 CHAPTER SUPPORT FOR PENSION REFORM THEBANK’S 3. Over 350 publications have addressed some aspect of pension reform. One hun- dred thirty one ESW studies have a regional or issue-specific focus; a number are dis- cussed inChapter 1as they provide an extension of the outline described in Strategy. Some studies were conducted as part of a country assistance strategy, while others were undertaken as research funded outside the operational process. 4. The categories were selected based o n prior knowledge of important pension issues and post-review categories that were found to be important. 5. The most ambitious study of income of t h e elderly i s by Edward Whitehouse (2000) who uses 1 1international studies to assess poverty among the aged. H e notes the im- portance of this type of analysis by stating that ”looking at pension replacement rates alone ignores other resources o n which the elderly can draw.” Whitehouse confirms conclusions about the lack of studies assessing the income of the aged by indicating ”We have also examined the World Bank’s poverty assessments. Few, however, pro- vide empirical evidence o n the economic status of the elderly.” This paper i s the first ina series o n poverty and income distribution issues in the design of old-age pension systems, but the series has not been completed. While the study covers 44 countries, t h e bulk of the analysis i s for the OECD. Data for ECA are from Grootaert and Braithwaite (World Bank 1998) and from the Luxemburg Income Survey. Data for L A C are from the IDB. 6. The ECSSD Working Paper No. 12, ”Older People in Transition Economies: An Overview of their Plight inthe ECA Region” (1999), is the perhaps most thematically comprehensive study of living conditions. Yet, even this report does not provide sat- isfactory data o n poverty and near poverty, or a n assessment of sources of income. These shortcomings reflect those of the poverty assessments that are the statistical nrecent papers, the Bank has been addressing this deficiency for basis of the inquiry. I L A C and AFR (see Bourguignon et. al, 2005 and Kakwani and Subbarao 2005. 7. Theoretical arguments have been made that coverage will increase i f contributions are tied more closely to benefits, but empirical follow-up has not evolved. 8. See Chapter 6 for a discussion of a lost opportunity to conduct ESW o n t h i s issue. The Pension Primer series (a compendium of Bank-commissioned papers o n pension reform issues) does contain some research o n these topics. 9. See Chapter 3 for details. 10. In Azerbaijan, the Kyrgyz Republic, and Moldova, the t h r u s t of Bank dialogue has been to discourage multi-pillar reforms i nview of economic and financial market constraints. Similarly, informal discussions between the Bank and the Russian au- thorities helped halt a premature multi-pillar reform, even though the Bank had pro- vided considerable funding to encourage the development of the reform. In contrast, the Bank was enthusiastic about multi-pillar reforms i nArmenia, Georgia, and Ukraine, and recommended them before economic and financial sector preconditions were i nplace. The Bank carried o n a n extensive policy dialogue with Poland and the Slovak Republic and both enacted multi-pillar reforms. However, discussions i nthe Czech Republic and Slovenia did not lead to any lending activity, and did not alter the path of locally developed pension programs. 11 .InUruguay, the Bank advised against the country’s o w n pension reform i nfavor of one more i nkeeping with Averting; as a result, the government turned to the Inter- American Development Bank (IDB) for assistance. The Bank supported the reform only after it was enacted. E l Salvador and Ecuador eventually used other advisors as well. 12. Despite a successful policy dialogue and initial Bank support, Mauritius has not implemented a multi-pillar system. Incontrast, the Bank was unsuccessful in dis- couraging a multi-pillar system inNigeria, where conditions are not favorable. How- ever, the Bank made premature efforts to move Zambia toward a multi-pillar reform despite its unfavorable conditions. 13. See Appendix B for a more detailed discussion of specific country reforms. 18 2 CHAPTER THEBANK'S SUPPORT FOR PENSION REFORM 14. There were some obvious exceptions, such as Brazil and Korea. 15. Over 70 percent of the Bank's pension lending was approved before t h e 2001 pub- lication of the Bank's official strategy on pensions. 16. identified only 18 operations that were 100 percent devoted to pension reform. These projects spanned all types of lending operations and were inEAJ? (China), ECA, and LAC. 17. See Chapter 4 for analysis on the initial quality of Bank projects. 18. These are Argentina, Hungary, Kazakhstan, Korea, Mexico, Peru, Russia, Turkey, Ukraine, and Uruguay. 19. Brazil, Korea, and Turkey did not. In fact, Brazil received the greatest total amount of pension support with Bank funding totaling $1.3 billion. 20. These figures excluded two outliers, Brazil and Korea. They included amounts spent o n both pillars. 21. The studies upon which t h i s review i s based are L A C case studies by Rofman, San Martino, and Valdes-Prieto (all forthcoming). 22. The World Bank also provided technical assistance to multi-pillar reforms in a number of other countries, includingCosta Rica, the Dominican Republic, and Nica- ragua, but funding was relatively minor. 23. See Appendix B for greater detail on reforms in specific countries. 24. Older participants can choose between the existing or old PAYG system and the funded tier, but their resulting placement in the funded tier i s binding. 25. Pension reforms in case study countries generally also had the participation of other international actors. Performance outcomes for this sample of countries were not related to the number or type of international actors involved. 26. See Appendix F for ratings of pension loans overall and for case study countries. 27. The development outcome of the pension component was identified by using a two-part rating of satisfactory or unsatisfactory. The development outcome for the project overall was taken from ICRs (self-evaluations by Bank teams) ICR reviews, and PPARs. The rating for the project overall i s based o n a six-part rating scheme, which was condensed to the two-part equivalent used i nthe pension component analysis. (The six-part project-rating scheme is: highly satisfactory, satisfactory, mod- erately satisfactory, moderately unsatisfactory, unsatisfactory, and highly unsatisfac- tory. The two-part equivalent scheme categorizes highly satisfactory, satisfactory and moderately satisfactory ratings as satisfactory; and moderately unsatisfactory, unsat- isfactory, highly Unsatisfactory ratings as unsatisfactory.) 28. Overall performance outcome is the subject of Chapter 4. 29. The management of projects containing pension components was spread across sector networks and boards, most frequently: Poverty Reduction and Economic Man- agement (Economic Policy and Public Sector); H u m a n Development Network (Social Protection); and Financial Sector. Five other sectors account for the remaining pro- jects: Human Development (Education and Health, Nutrition and Population); Transportation; Urban Development; Transport; and Energy and Mining. The latter two were more likely to be focused on pensions in the context of creating retirement packages in downsized industries. Pension components were included i nthree types of operations: adjustment lending, investment projects, and technical assistance. 30. AFR, EAP, MNA, and SAR were not included in the comparison because of small sample size. 19 3. Quality at Entry for Pension Reforms 3.1 The Bank's strategy on pension r e f o r m i s to support a multi- pillar framework as best practice if proper initial conditions are in place, w h i c h include (1)sound macroeconomic policies, (2) an ade- quate financial sector, and (3) implementation capacity. I f these condi- tions are not met, the Bank's strategy i s to improve the system through other means to create fiscal stability and protection o f the aged. Parametric reforms are recommended to complement multi- pillar reforms to improve the existing system or to precede the im- plementation of a multi-pillar system by helping to create the proper conditions. At the time of project design, the need for pension reform m u s t b e balanced against and/or coordinated with related social poli- cies to protect uncovered workers f r o m poverty in old age, as p a r t of the Bank's overall objective to reduce poverty among the aged. 3.2 Chapter 3 uses a set of indicators to assess whether the Bank exercised d u e diligence in ensuring that necessary conditions were m e t before supporting multi-pillar reforms and whether first-pillar assistance was considered w h e n countries did not implement multi- pillar systems. Whenever possible, these indicators are compared to the performance outcome evaluations of the Bank's assistance from the case studies. The chapter also assesses whether the Bank's objec- tive o f improvingthe welfare of the aged was addressed sufficiently a n d whether evidence substantiates the hypothesis that multi-pillar reform will increase the savings rate and enhance economic growth? 3.3 W h i l e the Bank did not advise funded pensions in many coun- n w h i c h initial conditions were unsuitable, the Bank acted too tries i quickly to support multi-pillar reforms i n other countries without ex- amining options for complementary safety-net programs to protect informal sector workers from poverty in old age. The Bank also sup- ported some reforms in w h i c h macroeconomic, financial sector, and institutional preconditions were not met, puttingthose reforms a t risk from the start. While many, but not all, ECA countries showed a readiness for reform, those in LAC were less l i k e l y to b e appropriate candidates. The development outcome ratings of the Bank's activities in case study countries in w h i c h preconditions were not m e t were of- ten rated unsatisfactory. 21 CHAPTER3 AT ENTRY QUALITY FOR PENSION REFORM When Were Only Single Pillars Considered? In nearly all 3.4 Very f e w countries i nw h i c h the Bank supported stand-alone countries in . single-pillar pension r e f o r m c o u l d have developed multi-pillar sys- which the Bank t e m . Invirtually a l l cases, these countries did not have satisfactory supported preconditions.2 Such single-pillar support was appropriate a n d in stand-alone concert with the Bank’s strategy. While multi-pillar reforms were dis- single-pillar cussed in a number of these countries, including China, India, Korea, reform, the a n d Turkey, the Bank did n o t insist that these reforms b e imple- countries did mented before economic prerequisites were i nplace a n d the political not have will to take such a step was firm. satisfactory 3.5 Brazil was by far the largest recipient o f Bank assistance for preconditions nw h i c h i t s budget defi- first-pillar r e f o r m in L a t i n America, at a time i for multimpillar cit was high a n d inflation rate excessive. The Bank also p r o v i d e d first- reform pillar assistance to Panama, a country with a stable macroeconomic environment a n d a strong financial sector, but one i nw h i c h pension coverage was l o w and the development o f a multi-pillar pension sys- t e m was n o t considered a priority. 3.6 The Bank also was cautious i n its activities in many poorer a n d less stable transition regimes, including most of Central Asia and the Caucasus. Both Turkmenistan and Uzbekistan h a d extremely weak barking systems, and Tajikistan h a d the lowest regional gross domestic product (GDP) per capital int e r m of purchasing p o w e r parity. The financial sector in Belarus was also underdeveloped a n d those i n the Caucasus have been borderline although regional im- provements have been taking place rather quickly over the past cou- nSlovenia were appropriate for a p l e o f years. W h i l e conditions i multi-pillar reform, i n the end, the authorities were not interested in pursuing this policy. This was also true of the Czech Republic, one of the m o r e successful transition countries, but one that h a d not sought m u c h Bank assistance overall. 3.7 Fourteen African countries received at least small amounts of Bank assistance for first pillar r e f o r m to restructure c i v i l service pen- sions and provident funds. The largest was Zambia, followed by Senegal. These countries are characterized by high poverty rates, poor financial sectors, a high p r o p o r t i o n o f foreign aid, and, in some cases, considerable government debt and inflation. N o n e would have been good candidates for multi-pillar reform, although the Bank discussed the option of multi-pillar r e f o r m with Cameroon, Cape Verde, M a u r i - tius, Senegal, and Zambia. The Bank did not always prioritize the Were Complementary Safety Nets Considered? need to expand safety 3.8 While the Bank refrained from pursuing multi-pillar systems nets to those outside in many countries with inadequate preconditions, it also often failed the formal system 22 CHAPTER 3 FOR PENSION REFORM QUALITY AT ENTRY t o prioritize the need for developing options for old-age safety nets outside the formal pension system in low-income a n d low-coverage countries. Out of eight low-coverage L a t i n American countries3 that enacted multi-pillar systems with W o r l d Bank support, only Bolivia created a comprehensive safety net, the Bonosol,4 inconjunction with its multi-pillar reform. Both Argentina5 a n d Brazil6 h a d r u r a l pro- grams that p r o v i d e d pensions to the aged, but b o t h h a d substantial weaknesses that the Bank failed to address i n its operations. 3.9 Many Central and Eastern Europe reformers, such as Latvia, ...nor always have targeted safety nets for uncovered workers, and the Bank pro- fully analyze vided assistance in a number of cases. But the Bank did not analyze the non- effectiveness of non-contributory options i ncountries like Albania, contributory Bosnia & Herzegovina, and the Kyrgyz Republic, where coverage i s options nAsia, Korea added a non-contributory emergency l o w or declining. I pension with World Bank support inthe context o f parametric reforms, but China has not addressed the issue of old-age r u r a l poverty even though the formal system covers only about 20 percent of the total population. With the exception of Mauritius, coverage rates i n African couniries where the Bank has h e l d discussions are less than 15 percent. I nZambia, w h i c h received signhcant World Bank funding to redesign its P A Y G system, neither the Bank nor the country undertook an analy- sis to identLfy needs and options, as feasible, for reducing poverty among the current and future uncovered elderly, an exercise w h i c h should have been conducted simultaneously with Bank funding for P A Y G redesign. Is the Economy Stable? Multi-pillar 3.10 Large fiscal deficits, stemming i np a r t from imbalances inpen- systems require sion revenues and expenditures, are often a motivation for countries a strong t o seek assistance from the World Bank i n reforming their pension macroeconomic systems. However, switching from a P A Y G system to a f u n d e d sys- environment t e m m a y n o t b e the best course for countries with fiscal imbalances that are d r i v e n by factors other than pension deficits. The transition costs of switching from a P A Y G system to a f u n d e d system will tem- porarily increase the fiscal deficit as the government m u s t continue to p a y pension benefits w h i l e some contributions are diverted into pri- vate funds. Countries should b e advised first to achieve fiscal sustain- ability through expenditure rationalization and revenue reform, in- cluding parametric reforms to their pension systems, before embarking o n a multi-pillar reform. Countries with high levels of public debt m a y n o t b e able to take on the additional debt derived from the start of funded systems. Countries whose budgets are heav- ily dependent o n external a i d also m a y not have sufficiently stable revenue bases t o support a multi-pillar system. 23 CHAPTER3 AT ENTRY QUALITY FOR PENSION REFORM 3.11 Stable monetary and fiscal policies are needed i f multi-pillar systems are to achieve long-run retirement-income objectives, as large macroeconomic imbalances, high inflation, a n d excessive debt bur- dens create uncertainty and destabilize financial markets. Further, high levels of government debt are l i k e l y to constrain the develop- m e n t of capital markets. Four indicators are used to evaluate macro- economic readiness for multi-pillar reform: the inflation rate, the cen- t r a l government budget balance, public debt, and the share of development assistance in total national income. Some Bank- 3.12 Inflation. The Bank supported multi-pillar reforms in a number supported multi- of countries with high inflation. High inflation was a p r o b l e m inL a t i n pillar reforms America in the early 1990s during the first round of multi-pillar re- were in forms (Figure 3). Ecuador, Peru, a n d U r u g u a y faced inflation rates of countries with over 35 percent at the start of their pension reform.7 E v e n after hyper- high inflation inflation subsided in ECA, price increases exceeded 15 percent in sev- levels eral countries instituting f u n d e d pillars. Fortunately, inflation de- clined thereafter, so that price increases were b e l o w yields on government bonds. Seven countries with high inflation at the time of r e f o r m were included in the case studies.* Of those, the development outcome o f the Bank’s assistance to Peru, Russia, and U r u g u a y was moderately unsatisfactory. W h i l e a number of countries m a y have postponed their reforms, the enactment of legislation was not pre- dicted on post-enactment decisions to delay. Figure 3. Many Countries Had High Inflation at Reform Estonia Bolivia Croatia Dominican Republic Costa Rica Argentina Nicaragua Poland Kazakhstan Hungary Russia Romania Colombia Mexico Latvia -5 5 15 25 35 45 55 65 75 a5 95 Percentage increase in CPI Source: World Development Indicators 2005, World Bank. 3.13 Fiscal balance. M o s t countries initiating multi-pillar pension re- forms h a d “moderate” fiscal deficits. Bolivia, Kazakhstan, Latvia, and Romania h a d budget deficits over 3 percent of GDP, a n indicator that 24 CHAPTER3 FOR PENSION REFORM AT ENTRY QUALITY fiscal conditions for implementation of a funded system were n o t ideal (Figure 4). Although Latvia's deficit was initially high, w h e n the f u n d e d tier was implemented i n 2001, the deficit h a d fallen to 1.4 per- cent of GDP. A s for Kazakhstan, a favorable prognosis for the energy sector was realized, p r o v i d i n g economic stability a n d increasing gov- ernment revenues. Bolivia, Latvia, a n d Kazakhstan were included i n the case studies. Of those, the development outcome o f the pension n Bolivia was rated moderately unsatisfactory.9 r e f o r m activities i Figure 4. Several Countries Had High Budget Deficits at the Time of Their Pension Reform Bulg .ia ifna -5.0 -4.0 -3.0 -2.0 -1.o 0.0 1.o 2.0 Overall budget deficit as YOof GDP Source: World Bank Indicators 2005, World Bank. 3.14 Governmenf Debt. U n d e r stable economic policies, govern- Governments ments m a y restructure expenditures, raise taxes, or use government need to finance debt to finance the transitional deficit from multi-pillar pension re- the transitional form.10 Countries i n L a t i n America with l o w e r coverage, ceteris pari- debt resulting bus, face less transitional debt than countries i n ECA as past pension from multi-pillar promises are smaller. Because the size of the f u n d e d pillar i s related to reform the size o f the transition deficit and i t s m e t h o d of funding, the Bank was prudent i n supporting larger PAYG pillars i n Central and Eastern European countries, a n d even pillars that were earnings a n d contribu- tion related. This support was consistent with the flexibility of pen- sion design proposed i n Sfrufegy. 25 CHAPTER3 AT ENTRY QUALITY FOR PENSION REFORM 3.15 Nonetheless, countries with l o w pension coverage can also have problems implementingfunded systems i f they have a l i m i t e d tax base and high levels of government debt. As a benchmark, public Countries with a debt of over 60 percent of GDP is generally regarded as extremely limited tax base high. Nicaragua and Bolivia both enacted Chilean-style reforms and and high levels received support f r o m the Bank. Both are HIPC countries, and their of government indebtedness will remain a problem for a multi-pillar pension reform debt may face even if some obligations are forgiven. Bolivia, a n case study country, bigger was rated moderately unsatisfactory o n overall development out- challenges come. Senegal and Uganda have recently been studying multi-pillar reforms, but they are also HIPC countries. So far, the Bank has n o t been sufficiently proactive in trying t o defer multi-pillar reforms i n highly indebted countries. 3.16 Development Assistance. Countries receiving a high proportion of Countries development assistance relative t o gross national income (GNI) are un- dependent on able t o function independently without donor resources and are n o t development fiscally independent. Further, if substantial development assistance is assistance are needed, poverty is likely to be high, and a poor population does n o t not good have sufficient income for discretionary saving through a pension sys- candidates for tem. InECA, although Georgia received substantial development assis- multi-pillar tance relative to GNI, the Bank s t i l l promoted a multi-pillar reform. reform Bank support t o FYR Macedonia was conditioned o n the implementa- tion of a multi-pillar reform even though it was heavily dependent u p o n donor assistance at 7.4 percent of GNI, but the Bank eventually supported the suspension of the reform. 1 1 3.17 T w o LAC countries that enacted multi-pillar reforms, Bolivia and Nicaragua, are heavily dependent u p o n donor assistance -at 9.0 and 13.6 percent of GNI, respectively.12 The Bank’s activities for the Bo- livian reform were rated moderately unsatisfactory inthe case study. Eventually the Bank helped place the Nicaraguan multi-pillar reform o n hold, but only after initially encouraging the reform. More recently, the Bank has supported Senegal, which has been considering multi- pillar reform despite having 9.3 percent of its GNI in development as- sistance funds. Is the Financial Sector Sound? Countries 3.18 According t o Vittas (1998), countries starting a funded system starting funded need “at least a small number of sound a n d well-functioning banks systems should a n d insurance companies coupled with a willingness t o implement have a well- capital market reforms and openness t o foreign expertise.” Impavido, functioning Musalem and Vittas (2001) also w a r n that systemic multi-pillar pen- financial sector sion reform i s unlikely t o succeed in countries inw h i c h the dominant banks are state-owned, financially insolvent, a n d operationally inept. 26 CHAPTER3 AT ENTRY QUALITY FOR PENSION REFORM 3.19 Banking Systems in Europe and Central Asia.l3 A number of ECA A number of countries that the World Bank assisted inmulti-pillar reforms h a d fi- ECA countries nancial sectors that did not have a sound financial system. At the time undergoing their pension reforms were enacted, four countries, Kazakhstan, Ro- Bank-assisted mania, Russia, and Ukraine, h a d financial sectors that, as evidenced by multi-pillar the EBRD’s financial system rating (Figure 5), did not exhibit: ( 1 )sub- reform had weak stantial progress inbank solvency, (2) a framework of prudential r e p - financial sectors lation and supervision, (3) full interest rate liberalization with little preferential access to cheap refinancing, (4) sigruficant lending to pri- vate enterprises, and (5) a sigruficant presence of private banks. 14 Three other countries, Bulgaria, Latvia, and FYR Macedonia m e t these criteria but did not have: (1)sigruficant movement inbanking laws to meet BIS standards, (2) well-functioning banking regulation and effective pru- dential supervision, (3) sigruficant term lending to private enterprises, and (4) substantial financial deepening. The banking systems of Croa- tia, Estonia, Hungary, Poland, and the Slovak Republic were effective according the EBRD standards. Figure 5. Poor Financial Sectors Characterize Some ECA Multi-pillar Reformers Hungary Slovak Republic Estonia Poland CroaCa FYR Macedonia Latvia Bulgaria Romania Ukraine Kazakhstan Russia 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 EBRD rating at time of reform Source: EBRD. 3.20 Countries that have n o t implemented multi-pillar pension sys- tems in ECA are more likely t o have weak financial sectors (Figure 6). The Bank encouraged a multi-sector r e f o r m in Georgia in 1996 and 1998 through adjustment loans, despite a weak financial sector.15 By con- trast, the Bank has been trying to discourage the Kyrgyz Republic from inaugurating a multi-pillar system. The Czech Republic and Slovenia, w h i c h have strong financial sectors, have exhibited little interest in multi-pillar reform. 27 CHAPTER3 FOR PENSION REFORM AT ENTRY QUALITY Figure 6. ECA Countries Without Multi-pillar Reform Are More Likely to Have Weak Financial Sectors Czech Republic I Slovenia Albania Basnia and Herzegovina Georgia Mddova Armenia Azeibaijan Kygyz Republic Tajikistan Belarus Uzbekistan Turkmenistan 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 EBRD Financial Rating, 2004 Source: EBRD. 3.21 Among multi-pillar reformers i nLAC, Argentina, Colombia, a n d Peru h a d financial sectors that m e t the minimumconditions for a multi-pillar reform. Their systems were characterized by declining margins, increasing levels of intermediation, a n d an increasing range of financial products and services. However, these countries also h a d a relatively high proportion of state ownership in the banking sector a t the time of their reforms. Some LAC 3.22 Among the LAC reformers enacting multi-pillar legislation af- countries ter 2000 -Costa Rica, the Dominican Republic, Ecuador, a n d Nicara- implementing gua -only Costa Rica h a d a w e l l developed financial sector at the funded systems t i m e of reform.16 While Bank support through policy dialogue, credits, lacked well- or,loans of these countries was considerably less t h a n that p r o v i d e d to functioning earlier reformers such as Peru and Argentina, the Bank did not dis- financial sectors courage the latter set of reforms because of unsatisfactory financial sector performance. The r e f o r m in Nicaragua was eventually put on hold, in line with Bank guidance, as subsequent analysis found the country unprepared for such an ambitious undertaking. 28 CHAPTER3 AT ENTRY QUALITY FOR PENSION REFORM 3.23 nLAC, the Bank i s currently w o r k i n g with Brazilian officials I to help the federal and state governments introduce a n e w and com- plementary DC pension fund for n e w c i v i l servants, providinga complementary f u n d e d benefit on top of the PAYG benefit. The Bra- zilian financial market i s satisfactory, and the equity market, with sigruficant occupational funds, i s better developed than in many countries with mandatory f u n d e d pillars. 3.24 Banking Systems in Other Regions. Except for Mauritius, no Af- rican country considering multi-pillar reform has a strong enough fi- nancial sector to support a multi-pillar reform.17 This i s consistent with the World Economic Forum’s (2000) ratings of confidence in fi- nancial services in southern Africa, inw h i c h only Botswana and South Africa h a d a higher rating than Mauritius. Based on best prac- tice, the Bank appropriately advised against a proposal for a multi- pillar r e f o r m in Nigeria where the financial sector i s characterized by high margins, low levels o f intermediation, and f e w financial prod- ucts or services. Despite the government’s inclination to adopt multi- pillar reform, the Bank considered supporting the Nigerian pension system i n a w a y that did not include multi-pillar reform. 3.25 InEast Asia, Korea’s financial markets are sufficiently devel- oped to support a multi-pillar reform.18 Incontrast, the Chinese finan- cial sector remains weak. Four state banks account for about two- thirds o f a l l deposits, lending primarily to state-owned enterprises a n d not to the booming private sector.19 While the Chinese have not yet started a multi-pillar system, China’s gradual approach to a l l re- form appears to have been instrumental in moving slowly, as the Bank’s advice did not adequately stress financial market readiness, but focused instead on actuarial projections and general information on reform. Can Implementation Be Effective? 3.26 One reason to shift to a privately f u n d e d pension system i s to Countries eliminate the potential for government interference. W h i l e f u n d e d pil- shifting to lars provide autonomy from government, they are vulnerable to cor- private pension ruption a n d weak governance within the private sector. A s a result, funds need to strong regulation and supervision of the pension funds, asset manag- ensure they ers a n d other financial intermediaries are essential if multi-pillar sys- have strong t e m are t o operate prudently and effectively. A l t h o u g h the presence regulatory of a private system can provide some balance against public power, a systems weak regulatory system m a y y i e l d insufficient protection for pension- ers in some countries, especially i nf u n d e d systems where partici- pants bear the r i s k of financial failure.20 InAfrica, Asia, and MNA provident funds and/ or partially f u n d e d PAYG pension plans have received below-market rates of return, resulting from high commis- 29 CHAPTER3 AT ENTRY QUALITY FOR PENSION REFORM sions, dubious investments, and outright theft, depending on the cir- cumstances. So far, there has not been documentation of a n y instances of fraud and abuse inmulti-pillar systems the Bank has supported.21 3.27 One signal of government commitment to regulatory r e f o r m i s the W o r l d Bank's index of control of corruption, w h i c h can b e used to assess the pofenfid for regulatory effectiveness. Even i f regulators are honest, they will b e hard-pressed to regulate financial assets i na country i nw h i c h business dealings are highly corrupt. The Bank sup- ported the enactment of legislation for f u n d e d pillars in 13 countries that h a d corruption ratings below the second quartile (50th percentile); five countries were i n the lowest quartile (25th) during the year of their reform (Figure 7).z Figure 7. Many Reformers Had Poor Corruption Index at the Time of Reform _____ costa Rica mgary nd .. .. .. .. .. .. . . .. .. .. .. . ........................... . .. ....................................... ................................................. . .... I FYR Madedonia 1 Colombia I El Salvador . . . . . 0 25 50 75 100 World Bank Institute "Control of Corruption'' percentile (closest year to reform) Source: Governance Matters IV: Governance Indicators for 1996-2004, D. Kaufrnann, A. Kraay and M. Mastruzzi (2005), World Bank. 3.28 The sustainability ratings of case studies were associated with corruption index ratings. InLatvia, where implementation was de- layed, the corruption index m o v e d into the positive range by the t i m e the funded pillar started. The multi-pillar systems i nRomania, Russia, and Ukraine are not yet fully i nplace, although the Bank has sup- n a l l three countries for ported systemic pension r e f o r m vigorously i many years. These countries a l l have negative corruption indices. While the Bank has h a d a lesser role in the Dominican Republic and Ecuador, where reforms were adopted i n 2003 a n d 2001, respectively, immediate steps are needed to determine whether effective regulation i s possible. 3.29 Regulatory problems are likely to arise amongfuture reformers as w e l l (Figure 8). Georgia, the Kyrgyz Republic, and Turkey all have ratings 30 CHAPTER 3 AT ENTRY QUALITY FOR PENSION REFORM of corruption below the second quartile. The Bank has assisted Turkey and Georgia intaking prelmunary steps toward implementing a multi-pillar reform, but not the Kyrgyz Republic. Among the potential African re- formers, only Mauritius has a corruption rating above the 50th percentile. Figure 8. Corruption Ratings Are Poor Among Some Potential Reforms I Mauritius Korea ~ w \ & x y China 7 ; Kyrgyz Republic h y lUganga k p qZambi? b5SqGeorgia w- h Kenya j -Cameroon 1 r I I , , 0 25 50 75 100 World Bank Institute "Control of Corruption" percentile (2002) Source: Governance Matters IV: Governance Indicators for 1996-2004. Are Higher Saving Rates Needed to Encourage Growth? 3.30 Strategy argues that the adoption of a multi-pillar system can The Bank increase national savings, and increased savings m a y i m p r o v e eco- should first n o m i c growth. 23 In developing a pension strategy, policymakers need assess whether to assess the need for i m p r o v e d savings. Inlow-savings countries, t h i s a country has a argument can b e compelling; but not so in high-savings countries. need for East Asian countries with high savings rates include China, Malaysia, increased Thailand, and Vietnam. If high savings rates are combined with low savings returns to assets, P A Y G pensions can provide better rates of r e t u r n than funded plans.24 Eight countries enacting funded reforms h a d savings rates in excess of 20 percent of GDP a t the time of their r e f o r m (Figure 9). China, Korea, Mauritius, and Turkey, a l l potential reform- ers, have savings rates of over 20 percent (Figure 10). 3.31 Low-income countries with negative savings rates also may Countries with prefer a PAYG pension system. Nicaragua and Cape Verde face nega- low savings may tive savings rates with external transfers supplementing domestic con- opt for a PAYG sumption. This m a y b e the situation in Senegal and Uganda as well, system where coverage is low and saving i s negligible. The reasons for nega- 31 CHAPTER 3 QUALITYAT ENTRY FOR PENSION REFORM tive saving should be fully understood before deciding on a multi-pillar reform, as forced savings for covered workers may b e inappropriate. Figure 9. Some Multi-pillar Countries Already Had High Savings Rates m arag mfSSSS33 FYR Macedonia V- L Bolivia t K K Bulgaria n\\\\\\\\\\\\u Kazakhstan - Dominican -15 -10 -5 - 1- 1 0 I 5 I 10 I 15 Percentage of GDP Source: World Development Indicators 2005, World Bank. 20 , lJ kraine I 25 Mexim Ecuador Hungary 30 I I 35 U d SS 40 32 CHAPTER 3 AT ENTRY QUALITY FOR PENSION REFORM Figure 10. Saving Rates Require Review Before Deciding on Multi-pillar Reform -20 -15 -10 -5 0 5 10 15 20 25 30 35 40 45 Percentage of GDP, 2002 Source: World Development Indicators 2005, World Bank. Summary and Conclusion 3.32 In a number o f countries, such as Peru a n d Nicaragua, the Bank did not assess the needs of the elderly before p r o v i d i n g support for proposed multi-pillar reforms. By contrast, inother countries, such as Bolivia and Latvia, considerable attention was p a i d to old-age pro- tection for those outside the formal system. Insome instances, multi- pillar reforms were supported in countries that did n o t need higher savings to stimulate development. These included Ecuador a n d Rus- sia, countries that also did n o t meet the Bank's standards for success- ful multi-pillar implementation based on macroeconomic and/ or fi- nancial sector readiness. 3.33 The Bank supported multi-pillar reforms i n quite a f e w coun- tries that m e t its fiscal a n d financial sector standards, including Croa- tia, Hungary, Latvia, and Poland, w h i c h h a d favorable macroeco- n o m i c a n d financial market preconditions a t the start. Their multi- p i l l a r systems are more likely to b e successful. The pace of r e f o r m in Latvia was particularly prudent, as the f u n d e d pillar was postponed until all the economic and financial preconditions were in place. 3.34 Nonetheless, quite a f e w countries started reforms without macroeconomic stability, banking sector readiness, moderate indebt- edness, and low-to-moderate levels of corruption. Russia a n d Ukraine s t i l l have weak financial sectors and ratings b e l o w the 50th percentile for control of corruption. FYR Macedonia receives a high proportion of development aid. InL a t i n America, multi-pillar reforms were en- acted in a number of countries with weak financial sectors, including Bolivia and Mexico, and four later reformers - Costa Rica, the Do- 33 3 CHAPTER FOR PENSION REFORM QUALITY AT ENTRY minican Republic, Ecuador, and Nicaragua. Most of these countries are characterized by poor ratings of corruption control b e l o w the sec- o n d quartile. 1 .Implementation readiness is addressed further i nChapter 5. The sample of coun- NOTES nt h i s chapter’s evaluation consists of all countries that have under- tries included i taken systemic reforms or are prospective reformers, and that have sufficient docu- mentation for analysis. For instance, some countries, such as India, with which t h e Bank has had a substantial policy dialogue, are not included, as n o formal project has gone to Board. Other countries, identified in interviews, are not included due to a lack of documentation. As noted inChapter 2 , it i s difficult to assess the impact of the Bank’s policy dialogue i nthese circumstances. 2. See Appendix B for a summary of Bank operations and Appendix E for indicators of economic and financial sectm conditions. 3. Low-coverage is defined as coverage of less than 30 percent of the working popula- tion. 4. The Bonosol provides a benefit for all those who were 21 years old as of 1995, so it i s comprehensive for all elderly over the age of 65 until2039. 5. InArgentina, the case study notes that the decrease incoverage h a d been visible for several years, but the Bank only recently started to consider it, and Bank documents have not yet recommended relevant policy. The national non-contributory scheme nreducing poverty among the aged, and the schemes de- was relatively ineffective i veloped by provinces also had a limited scope, since they were arbitrarily assigned, given budgetary restrictions, and candidates accumulated o n long waiting lists. Bank assistance helped reduce some laxity and heterogeneity between provincial schemes, but it did not help reduce inequities i nthe system. 6. InBrazil, the case study finds that the proportion of the rural o l d with n o income from work or pensions i s lower than both the urban old and rural young i n1998. ESW i n1995 found that the level of the rural pension was above the poverty line, and it proposed that the rural pension be scaled back, but this recommendation was not adopted. ESW also proposed some modest targeting in the rural sector, but the objec- tives relating to the old poor were excluded from the five proposed priority goals i n the Structural Adjustment Loans even though they would have greatly increased the fiscal efficiency of the program. The exclusion of improvements to the rural poor was a reflection of the mentality that bringing up fiscal issues would lead to a c u t inbene- fits for the old poor. 7. In the medium term, post-reform, Argentina and Uruguay suffered economic cri- ses. 8. These were Hungary, Kazakhstan, Latvia, Mexico, Peru, Russia, and Uruguay. 9. The development outcomes according to the case studies were satisfactory for both Latvia and Kazakhstan. 10. See Appendix D for a discussion of the fiscal and debt implications of funded pen- sion reform. 11.The case study rating of FYR Macedonia is based o n the decision to place the multi-pillar reform o n hold; it is critical of the initial attempt to implement the reform. 12. While Nicaragua’s reform was put o n hold, the decision to proceed i nthe first place was not well advised given its significant reliance o n donor assistance. 13. The framework for measuring financial sector quality is based o n EBRD ratings, in- ternal Bank assessments, and a more comprehensive analysis of the financial conditions of each country from the internal FSAP assessments. 34 CHAPTER3 FOR PENSION REFORM AT ENTRY QUALITY 14. These EBRD measures of financial sector strength roughly parallel the Bank’s own measures of the financial sector. 15. This is confirmed by the Bank’s analysis. 16. In Costa Rica, while the banking system continues to be dominated by public banks and a Bank assessment of the financial sector was less than satisfactory, the FSAP analy- sis of the financial sector indicates that it is reasonably well-developed. Inthe Domini- c a n Republic, a Bank-Fund assessment pointed to signhcant issues surrounding the soundness of the financial system. 17. In Cameroon, banks remain vulnerable to fundamental liquidity r i s k and credit nthe trade balance. While the banking system i r i s k as a result of large movements i n Senegal is generally well regulated, it is still vulnerable to government pricing poli- cies, although the government is n o longer managing the day-to-day affairs of banks in which it holds a stake. The banking sector i nUganda has improved, but on-site ex- aminations continue to identdy sigruficant under-provisioning, and indicate that capital is understated. 18. However, Korea already has a high level of national savings, which may deter the government from adopting a multi-pillar system. 19. Poorly defined property rights and reports of corruption and misappropriation sug- gest that the central government may find it difficult to distinguishbetween what it owns and what is owed to it. I f so, the banks’ non-performing loans could jeopardize central government finances (China CAE 2004). 20. Strategy did n o t investigate the relationship between a culture of corruption and effective regulation of the private sector in its discussion of political risk and power. 21. However, this could simply be due to a lack of documentation, as countries with l o w incidences of corruption have well-documented instances of fraud and abuse. 22. Since the World Bank’s governance database provides information for 1996,1998, 2000,2002, and 2004, the data used are for the year closest to the reform. For earlier re- formers, 1996 is used and for later reformers, 2002 i s used. 23. Strategy states ”funded systems hold some advantages over traditional systems. The second pillar would be able to provide better income replacement for a given contribution rate.. .enhance national savings, promote capital market development, and reduce labor market distortions by linking contributions to benefits.” Perspective states, under a multi-pillar system, ”The saving rate and consequently the level of capital accumulated and output produced may be higher.. .Higher savings rates are associated with technological changes of the capital dependent type leading also to a higher growth path.” Inaddition, Perspective cites Feldstein (1996) in arguing that consequently ”the marginal product of capital exceeds the market rate of interest-as capital markets are never fully integrated-creating another gain for the national economy from a funded scheme.” na PAYG system is equal to the rate of growth of covered av- 24. The rate of return i erage wages plus the rate of growth of the labor force. 35 4. The Impact of Pension Reforms 4.1 Pension reforms require decades of implementation before a complete evaluation i s possible. For most reforming countries, too lit- tle time has elapsed t o evaluate the outcomes of Bank-supported re- forms, but indirect indicators can b e used to gain insight i n t o the de- velopment outcomes. 4.2 The first section uses such indicators to evaluate whether old- age income security has been achieved. While only implicit in Strut- egy, Averting a n d Perspective indicate that, w h e n appropriately imple- mented, multi-pillar systems ought to offer greater retirement income security than P A Y G systems by: (1)earning higher rates o f r e t u r n from diversified investments, and (2) spreading political and systemic risk between the public and private sectors. I norder for the pension system to provide retirement income security, the system also has to b e fiscally sustainable. The second section reviews improvements in the financial balance of PAYG systems a n d the transitional deficit re- sulting from multi-pillar reforms. 4.3 The third section reviews whether secondary objectives from multi-pillar reforms were achieved, including increased national sav- ings, capital market development, and labor market flexibility.' Fi- nally, the last section reviews the Bank's activities from the in-depth case studies to provide a full assessment of the development out- comes within countries, taking the entire r e f o r m process into account, rather than each indicator individually. 4.4 Based on existing evidence, multi-pillar reforms i n many countries, as implemented, have n o t i m p r o v e d old-age income secu- rity. Inmany cases, investments i nthe mandatory f u n d e d pillar are not w e l l diversified and instead are concentrated in government bonds. While in many cases the government bonds offer high rates of return, the high returns often reflect high levels of macroeconomic a n d investment risk. 4.5 Evidence o n savings and capital market formation i s also mixed. But this could b e in part a result of poor fiscal policy undermin- ing the potential for better outcomes. Lastly, linkages between contri- butions and benefits do not appear to have improved the efficiency or formalization of labor markets, as coverage has remained stagnant in many countries. 37 CHAPTER4 THEIMPACT OF PENSION REFORMS Income Security Outcomes 4.6 Has Risk Been Diversifid? Multi-pillar pension plans are ex- Private pension pected t o h o l d a variety o f securities, including higher-yielding equities funds need to and lower-risk international assets? but many pension portfolios are diversify risk primarily invested i n government debt. 3 Undiversified portfolios are a through varied result of government-imposed investment guidelines, a lack of domes- investments tic investment opportunities, and/or economic crisis.4 W h i l e the expan- sion of the demand for government debt can improve market efficiency and lead t o the development of longer-duration instruments, a concen- trated investment strategy in government debt fails to y i e l d the benefits provided by diversification? From the point of v i e w of macroeconomic policy, pensions invested primarily in g o v e m e n t bonds are little dif- ferent f r o m PAYG systems.6 Many private 4.7 InLAC, except for Chile and Peru, pension portfolios for re- forms supported by the Bank are heavily concentrated in government funds in LAC securities (Table 4).7 Colombia and Uruguay have a m o r e diversified are portfolio than other LAC countries but s t i l l maintain a majority share in concentrated in government securities. Insome countries, the domestic bond market government was not developed at the time o f multi-pillar reform. Ironically, the di- bonds versification of Peru’s portfolio partly results from poor debt manage- ment. During the early 1990s, there was no domestic bond market in Peru, and the government h a d to borrow entirely offshore and use a d hoc financing measures for its spending. 8 The combination of tight in- vestment guidelines and the still illiquid market for government debt has l e d t o greater portfolio diversification into capital markets. Table 4. In LAC, Only Some Funded Pension Portfolios Are Well-Diversified (percentage of holdings as of December 2002) Government Financial Corporate Investment Foreign securities institutions bonds Equities funds securities Other Argentina 76.7 2.6 1.I 6.5 1.8 8.9 2.4 Bolivia 69.1 14.7 13.4 0.0 0.0 1.3 1.5 Chile 30.0 34.2 7.2 9.9 2.5 16.2 0.1 Colombia 49.4 26.6 16.6 2.9 0.0 4.5 0.0 Mexico 83.1 2.1 14.8 0.0 0.0 0.0 0.0 Peru 13.0 33.2 13.1 31.2 0.8 7.2 1.5 Uruguay 55.5 39.6 4.3 0.0 0.0 0.0 0.5 Source: Keeping the Promise, based on data from AIOS, FlAP (data for Colombia). Note: Information for Colombia refers only to the mandatory pension fund system. ECA funds are 4.8 Diversification i n ECA i s also limited. H u n g a r i a n pension not yet well- funds invested roughly 70 percent of their assets in government diversified bonds in mid-2003, only somewhat l o w e r t han the 80 percent share a t inception. W h i l e they are slightly more diversified, Polish and Croa- t i a n funds are s t i l l invested a t roughly 60 percent in government 38 CHAPTER4 THE IMPACT OF PENSION REFORMS bonds in 2002.9 Kazakhstan pension funds have diversified consid- erably since 1998, w h e n the r e f o r m began (Figure 11).This diversifica- tion, in part, has been a result of a donor-assisted drive to develop n e w financial sector instruments. 4.9 Are Rates of Return Higher? The long-term stability o f the returns to funded pillars i s difficult t o evaluate due to sovereign credit risk, capital market volatility, and management costs. Gross rates of return It is difficult to for the funded pillars of many multi-pillar systems, however, have determine been favorable compared to wage growth-a proxy for the implicit rate whether private of return t o the P A Y G system (Table 5).’0 Ina l l countries except Ka- funds will enjoy zakhstan, wage growth since the start of the multi-pillar system has high rates of been less than the rate of return on assets. But gross rates of r e t u r n do return in the not account for administrative costs of managing the funds. I f a reduc- long run tion of 1.5 percent i s assumed for managing the funds (a fee structure found in some countries), then the positive wage-interest rate differen- tial is reduced t o around 2.5 percentage points for Peru and Poland.ll Figure 11. Pension Funds Have Become More Diversified in Kazakhstan Portfolio distribution in 1999 Portfolio distribution in 2002 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% NSAFs SAF Total NSAFs SAF Total Type of pension fund Type of pension fund Comme!ical bank deposits, Commerical bank deposits li3 Domesbc corporate securibes Securifes of international financial organizafions - rn Kazakhstan state securifes Foreign, securities Domesbc corporate securifes w Kazakhstan state securibes Source: National Bank of Kazakhstan. Note: NSAF- non-state accumulation fund; SAF - state accumulation fund. 4.10 Inmany countries, high returns are a result of high interest o n government bonds and, as such, are related to the risk profile of the country’s sovereign debt. Among the countries listed inTable 5, the 39 CHAPTER4 THEIMPACT OF PENSION REFORMS High returns in 1999 Standard & Poor's Sovereign Debt Rating for Chile w a s the , private funds are highest, a t A-, the 6th highest category out o f 18 alphabetical ratingsJ2 sometimes the Poland h a d the next highest rating at BBB, the 8th of 18, and Kazakh- result of investing stan was ranked the lowest, a t BB, at the 13th. Sixty-nine countries in high-risk were included i n the ratings. only 5 countries were rated l o w e r than government debt Kazakhstan; 26 were rated higher than Chile. 4.11 Funded pensions are proposed to reduce the demographic and Private funds are political risks of P A Y G systems. But pensions from f u n d e d pillars, sensitive to capital even in countries with mature, well-regulated, and highly efficient market fluctuations capital markets are subject to other risks, such as capital market vola- tility, w h i c h can lead to fluctuations in replacement rates across co- horts, depending o n the conditions of the market at the t i m e of labor- force entry until retirement. For example, with equity investments, pensions will be quite sensitive to the exact year o f retirement. 13 E v e n with investments in government bonds, replacement rates can vary considerably i f the authorities d o not predict long-term trends in re- turns accurately w h e n they set contribution rates for the f u n d e d p a r t of the pension systemJ4 Table 5. Real Returns Have Outpaced Wage Growth for Funded Pensions in Most Countries Difference Real rate of Real wage between ~ S8P Sovereign return growth rate of return Debt Rating Country (since reform) (since reform) and wage growth (1999) Argentina 11.7 -0.8 12.5 BB Bolivia 16.2 8.8 7.4 BB- Chile 10.5 1.8 8.7 A- Colombia 11.8 1.4 10.4 BB El Salvador 11.3 -0.2 11.5 BB+ Mexico 10.6 0 10.6 BB Peru 5.7 1.8 3.9 BB Uruguay 9.5 3.6 5.9 BBB- Poland 7.5 3.5 4.1 BBB Kazakhstan 5.8 8.4 -2.6 B+ Source: FlAP (2003) and Hammer, Kogan, & LeJeune (2004) "Modeling Country Risk Ratings Using Partial Orders". Note: Based on a subset of countries with available data. The S&P debt ratings ranged from AAA for many OECD countries to CC for Pakistan on a scale of 18 ratings. 4.12 Government interference.One u n d e r l y i n g motivation for a multi-pillar system i s to limit government interference in retirement income security. But Bank-supported pension reforms have not al- ways been effective in controlling government interference, especially during economic crises. For instance, during the Argentinean crisis, the government forced pension funds to take up government debt.15 Similarly, following the Russian crisis, the Kazakhstan government 40 CHAPTER4 THEIMPACT OF PENSION REFORMS "encouraged" pension funds to exchange government bonds h e l d be- Private funds are fore the currency devaluation for n e w issues, effectively reducing the not necessarily rate o f return. Government interference took place in Bolivia by forc- protected from ing pension plans to accept Bonosol bonds.l6 government interference Fiscal Sustainability 4.13 Fiscal sustainability i s a long-term target related to the extent to w h i c h parametric reforms have reduced unfunded pension liabili- Fiscal ties and the degree to w h i c h the funding of transition costs o f multi- sustainability is a pillar reforms has been fiscally responsible.17 A thorough analysis of long-term objective the financial sustainability of pension r e f o r m requires a n evaluation of multi-pillar of the actuarial projections of the PAYG system made before and after reforms the reforms. Equally important t o fiscal sustainability i s the fiscal stance independent from the pension system a n d associated transition cost (Figure 12). Figure 12. Fiscal Deficits Have Grown in Many Countries with Second Pillars 10.0 9.0 8.0 7.0 $ 6.0 Lc s 5.0 .- v) J 4.0 0 E i p" 3.0 2.0 1.o 0.0 -1 .o Years since pension reform +ArgenSna +Bolivia +Colombia &Uruguay Source: World Development Indicators 2005, World Bank. 4.14 The W o r l d Bank assisted many countries i nimplementing pa- rametric PAYG reforms to strengthen their system's fiscal balance. With Bank support, Brazil i m p r o v e d its fiscal position sigruficantly over 1998-2003, allowing it to survive two serious lapses ininvestor confidence. The cuts i npension expenditure achieved i nthe reform, 41 CHAPTER4 THEIMPACT OF PENSION REFORMS Bank assistance w h i l e temporary, were a positive step toward more permanent stabil- has helped ity. The Bank’s assistance to the Kyrgyz Republic reduced pension costs improve short- by one-half as a percentage o f GDP by 2002, compared to 1995. The re- term fiscal form helped balance the budget by 2003. I nKazakhstan, the fiscal defi- sustainability cit declined as a percentage of GDP, but was bolstered by the receipt of sigmficant oil revenues, an event exogenous to pension policy. In some LAC 4.15 na number of Latin American countries, including Argentina, I countries, Bolivia, Mexico, Peru, and Uruguay, insufficient parametric reforms i n insufficient c i v i l service pensions and other P A Y G plans created unsustainable parametric pension deficits.’* InArgentina and Bolivia payouts increased more reform has led than expected due to fraudulent claims and a lax interpretation o f rules to unsustainable (Ramachandran and Kissedes 2005). I nECA, the private sector evaded pension deficits b o t h first and second pillar contributions w h e n tax rates were high. I n these countries, the Bank focused more on developing multi-pillar sys- t e m than on implementing the complementary parametric reforms. 4.16 The World Bank did n o t make actuarial projections of the fis- Minimum cal expenditures required to meet minimum guarantees p r o v i d e d by guarantees may some multi-pillar system for pensioners whose calculated pensions, worsen the based on their accumulated assets, would b e lower t h a n legislated fiscal situation minimums. This is an issue i n a number of countries i n ECA and LAC. Minimumguarantees and intermittent coverage will lead to higher general revenue expenditures whenever workers f a i l to accu- mulate sufficient resources i ntheir i n d i v i d u a l accounts beyond the minimum pension. Projections are needed to assess whether this p r o b l e m i s imminent, especially inChile and Kazakhstan. I naddition, pensions i n multi-pillar systems can gradually f a l l b e l o w the mini- mum if the guarantee i s indexed a n d the full pension i s not, w h i c h m a y pose a p r o b l e m for Poland. Savings and Capital Market Development The effect of multi-pillar 4.17 Have Savings Rates Increased? The impact of multi-pillar r e f o r m reforms on on savings i s inconclusive (Figure 13). Savings rates in Kazakhstan, savings is Latvia, and Peru increased after multi-pillar reforms, but i n Kazakh- unclear stan, growth i n oil revenue i s m o r e l i k e l y than pension r e f o r m to have i m p r o v e d gross domestic savings. By contrast, savings rates i nBolivia a n d U r u g u a y remained virtually unchanged, w h i l e rates i nColombia, Hungary, and Mexico declined. Argentina, Colombia, and Peru still have savings rates of less than 20 percent of GDP, a n d Bolivia and Uruguay have savings rates under 15 percent.19 4.18 The relationship between pension r e f o r m and savings i s complex, depending o n the w a y i nw h i c h the fiscal deficit has been financed, the reaction o f financial markets, and the reactions of work- ers? InKazakhstan, the fiscal deficit declined as a percentage of GDP, a n d the savings rate increased. InColombia, Hungary, and Uruguay, 42 CHAPTER 4 THE IMPACT OF PENSION REFORMS savings rates are highly correlated with fiscal deficits, suggesting that Pension reform poor fiscal policies could reduce possible positive gains to savings and savings are from multi-pillar reforms. Savings and the fiscal deficit have been un- not always correlated in Peru, however, where savings increased. correlated Figure 13. Savings Rates Increased Only in Kazakhstan Trends in Gross Domestic Savings ( % o f GDP) 30 25 . ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... . n n a 6 20 s m a g, 15 .- r: > 2 10 5 0 1 2 3 4 5 6 7 8 Years since start of reform +Argentina +Bolivia -k Colombia +Hungary ,+Kazakhstan -0- Mexico +Peru +Uruguay Source: World Development Indicators 2005, World Bank. Note: The most recent data for Argentina was omitted as it likely resulted from a reduction in GDP rather than an increase in saving. 4.19 Have Capital Markets Developed? So far, most capital markets Multi-pillar have not developed sigruficantly as a result of multi-pillar pension re- reforms has not form, even w h e n financial sector and capital market regulation has had a strong improved. While government bonds lengthened their maturities in impact on some countries i nconcert with the needs of pensioners and pension capital market plans, this has been an inadequate substitute for diversification, par- development n countries v i e w e d as poor sovereign credit risks. Nor did ticularly i pension plans diversify towards corporate securities or bank deposits, both capital market alternatives with w h i c h to finance business ex- pansion and development. 4.20 Initially, hopes for pension-stimulated capital market devel- This may be opment relied on equity-market development, although this emphasis attributed to the has shifted in recent years. The equity market impact of multi-pillar predominance r e f o r m has not been strong (Figure 14).21 But l i k e savings, equity mar- of government kets are influenced by m y r i a d of unrelated factors. Still, equity market bonds in development has not proceeded in Colombia, or H u n g a r y although pension there has been growth in Peru. Colombia has h a d a l i m i t e d a n d frag- portfolios mented equity market for years, with trading concentrated in 10 43 CHAPTER4 THEIMPACT OF PENSION REFORMS stocks and no n e w issues. The Mexican equity market declined in the post-reform p e r i o d from being one of the strongest in L a t i n America. While equity markets expanded in Peru, they accounted for only 25 percent of GDP after nearly a decade of reform.22 Figure 14. Market Capitalization Remains Quite Low - . 55 50 45 40 n 35 n 2 .-30 0 25 n t 20 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 Years since reform +Colombia 6Hungary +Mexico +Peru Source: Internal Bank data. Note: Kazakhstan is not included as market data were extremely unstable, no doubt reflecting a very small market, and possibly, changes in oil stock listings. Economic 4.21 Of course, with sigruficant shares o f pension portfolios in gov- conditions also ernment bonds, equity markets are unlikely to grow as a result of influence asset pension reform. An important question i s whether the lack of equity allocation market development i s a cause or a result of the concentration of in- vestments in government bonds. The development of equity markets m a y require m o r e than a potential pool of funds, i f companies are not ready to capitalize by issuing shares on the open market. For example, as the b a n k i n g system became stronger in Hungary, Banks began to provide capital for expanding businesses through loans, reducing the need for companies to go to the equity market. 4.22 Economic conditions play a role in asset allocation as well. Countries that h a d high fiscal deficits, such as Colombia and Hun- gary, are u n l i k e l y to experience capital market expansion. Trends in asset allocation i nArgentina illustrate how financial crises can affect multi-pillar pension systems. Argentina’s pension f u n d s h a d been be- coming m o r e diversified until 1998, w h e n the economy entered into a depression that eventually l e d to a crisis in 2001 (Figure 15). After 44 CHAPTER4 THE IMPACT OF PENSION REFORMS that, government bonds took up the lion’s share of investments, i n part, because of government pressure. Figure 15. Post-Crisis Pension Portfolios in Argentina Fled to Government Bonds 80 P en ........................................................ \ ................ 1994 1995 1996 1997 1998 1999 2000 2001 2002 +Govern ment debt +Other assets Source: Argentina Case Study. 4.23 Countries running a current account surplus c o u l d have di- versified by investment i nforeign markets, but most developing countries run current account deficits (Ramachandran a n d Kissedes 2005). Inaddition, in small capital markets, such as that of Peru, pen- sion funds would affect the prices of assets and m a y have to invest heavily inbank debt instead (Hanson and Ramachandran 2005). The Formalization of Labor Markets The impact of 4.24 Strategy argues that a multi-pillar system should ”interfere less multi-pillar reforms ni n d i v i d u a l labor supply a n d saving decisions.” One measure of a i on ECA’s labor more efficient labor market would b e the degree o f formalization, as market efficiency reflected in the pension coverage rate, since pension coverage i s o n l y is inconclusive important in the f o r m a l economy. Infact, multi-pillar pension reforms supported by the World Bank have not achieved higher participation (Figure 16).z 4.25 Post-pension r e f o r m participation increased somewhat in Chile a n d Colombia, but plummeted in Argentina. Small gains i n Bo- l i v i a are associated with the pension reform. One explanation for the stagnation of coverage rates c o u l d be the presence of minimumguar- antees, w h i c h m a y encourage low-income workers to limit their years o f contributions i nthe f o r m a l system. Others are high contribution 45 CHAPTER4 THEIMPACT OF PENSION REFORMS rates and lack of economic growth. Whether this is because workers mistrust social security or simply evade t h e payroll tax i s unclear.24 4.26 The impact of the Bank's support of multi-pillar schemes o n la- bor market efficiency inECA is still unclear. As a result of the socialist legacy of virtually full formal-sector employment, ECA pension cover- age rates are considerably higher than those of countries at similar lev- els of per capita income (adjusted for purchasing power parity). Latvia has witnessed greater labor market formalization, however, with pen- sion coverage increasing by about 3 percent between 1995 and 1999 af- ter the pension reform became effective but before the funded system was introduced i n2001. InKazakhstan, despite a relatively large in- formal sector, the number of covered workers increased since 1998, but this is likely to be related t o renewed wage gains and growth stimu- lated by o i l revenues. By contrast, the number of contributors to the Hungarian pension system remained virtually constant. Figure 16. Pension Participation Rates Have Not Changed in LAC 0.0 4 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 - + - A r gSalvador enha - I --Bolivia +Chile - + -Colombia - -A- - Ecuador +El --o-Mexico +Nicaragua +Uruguay +Brazil +Cos$ Rica Source: Keeping the Promise of Social Security in Latin America, Chapter V; Figure 5.4, p. 100. Development Outcomes of World Bank Assistance Development 4.27 commissioned independent consultants to conduct case studies outcomes for that w o u l d provide a more comprehensive look at the Bank's involve- case study ment i npension program activities over time (Table 6, page 49). The countries do not ratings indicate a range of success regarding the influence of the Bank's show a clear activities o n development outcomes and expectations that these re- trend forms will be sustainable i nthe future. InArgentina and Uruguay, the Bank was perceived as pushing reforms that were n o t adapted to coun- 46 CHAPTER4 THEIMPACT OF PENSION REFORMS try conditions. InArgentina, Bolivia, Brazil, and Mexico, insufficient as- sistance or attention was placed o n long-run projections. The highly satisfactory rating for Bulgaria reflects the Bank’s concern for a l l the factors involved i na reform, including financial sector readiness, politi- cal will, and institutional readiness. The common themes i nthe satis- factory ratings for the Bank’s activities for ECA and East Asian coun- tries are a focused reform agenda (albeit with some failures i n implementation i nLatvia and Hungary) and good operational advice, (albeit with little input into policy i nKazakhstan). The use of experts exposed Korea t o international best practice and the effectiveness of policy advice i nChina contributed t o the success of its reforms. The Bank gave sound advice to the Kyrgyz Republic by advising against the institution of a multi-pillar system inthe absence of appropriate fis- cal and financial sector preconditions. 4.28 No easy formula explains why the impact of the Bank’s activi- ties on development outcomes i n some case study countries i s better tha nin others. W o r l d Bank support has varied by l o a n amounts, in- vestment-adjustment mix, a n d funding for project preparation. Bul- garia h a d relatively f e w f o r m a l ESW studies but the Bank‘s contribu- tion to development outcome was highly satisfactory. Russia h a d a dozen ESW studies and the Bank’s contribution to development out- come was rated moderately unsatisfactory. With only two loans and credits, the Bank’s assistance t o China’s development outcome was satisfactory, as was Brazil’s n i n e loans and credits. Latvia received only $26 million i nloans a n d the Bank‘s assistance to development outcome was satisfactory, compared to Peru, w h i c h received $364 mil- lion and a development outcome of moderately unsatisfactory. Summary and Conclusions 4.29 A positive impact of the Bank‘s activities on development out- comes for multi-pillar pension reforms requires a combination of meas- ures w o r k i n g together, including effective fiscal policy, cost-reducing parametric reforms, and encouragement o f capital markets.25 The full effect of the Bank’s activities on pension reform on development out- comes can only b e evaluated if the program is assessed i nits entirety, in combination with a set of indicators, rather than by the performance of the pension components of individual loans or even groups o f loans. 4.30 The outcome of the Bank assistance to multi-pillar reform falls short of achieving the objectives identified i n Strategy. To some extent, this reflects the short time since the inception o f reform. But, pension portfolios i nmany countries are concentrated i ngovernment securities. Only the Chilean, Colombian, and Peruvian pension portfolios are rela- tively w e l l diversified. While parametric and multi-pillar pension re- forms have improved the financial balance of P A Y G systems, addi- 47 CHAPTER4 THEIMPACT OF PENSION REFORMS tional reforms are often needed. Multi-pillar systems remain open to political influence, especially intimes of economic crisis. Multi-pillar pension reforms have n o t yet increased savings or substantially devel- oped capital markets. High rates of interest o n government bonds and regulatory limits o n domestic equity investments may have stifled capi- t a l market expansion. By contrast, countries with low coverage need to expand their safety nets to improve the welfare of the elderly by other means, such as non-contributory options. 48 8 co e In I NOTES 1 .Very little research has been conducted o n the direct impact of pensions o n eco- nomic growth. One recent study (Davis and Hu 2004) found a positive relationship between pension assets, growth, and capital for emerging market economies. How- ever, the countries with the highest pension assets as a proportion of GDP were Ma- laysia, Fiji and Chile, two of which do not have a multi-pillar system and the last, Chile, exhibits more satisfactory performance i nmany areas compared to other multi- pillar countries inLatin America. Theses issues will need to be addressed further in the future as multi-pillar reforms continue to grow and asset holdings expand. 2. While investing abroad will help divers* t h e portfolio and decrease investment risk, especially when the domestic capital market is undeveloped, it exposes the port- folio to exchange rate volatility and currency risk. 3. Some countries, such as Estonia, have invested in government bonds of different countries. The r i s k profile of s u c h an investment strategy would be quite different from undiversified investments i na country’s o w n sovereign debt. 4. Duringtimes of economic crisis, pension fund portfolios may naturally shift to- ward government debt when other assets lose value or become more risky. 5. Nonetheless, diversification to highly risky private equity or debt instruments nassets below in- would not improve the r i s k profile of participants. Investment i vestment grade should be precluded by effective regulatory controls. 6. A number of proponents of funding, however, would say that participants may re- ceive a stronger guarantee if their pensions are funded through bonds than if they are funded o n a PAYG basis. 7. Inmost countries with funded pension systems, portfolio ceilings are set by regula- tion rather than applying the prudent investor rule. Peru i s the only country in which government securities c a n be a maximum of 30 percent of a pension f u n d s portfolio. Inother countries, funds can invest from 50 percent to 100 percent of their assets in government issues. Inmost countries, there are very l o w limits o n the percentage of nforeign securities. Only Bolivia permits up to 50 percent assets that can be invested i of assets abroad. 8. The absence of a well-developed domestic market for government debt also weak- ened bank risk management. 9. The share of government bonds i nEstonia was quite l o w at 34 percent of the port- folio, but Estonia’s reform was not among those supported by the Bank through loans or credits. 10. The International Labor Organization estimates that the rate of return i nPoland was less than inflation from December 1999 to 2004 and that the situation i s similar i n Hungary. 11.This difference could decline over time if economies of scale in pension manage- nreducing costs of man- ment are achieved and regulatory agencies are effective i agement and marketing. See Dobronogov and Murthi2005. 12. The year 1999 is an intermediate marker for the progress of the various pension reforms. 13. There are many other ways to reduce investment risk, including the use of rela- tively sophisticated approaches such as hedge funds. 14. See Thompson (1998) for simulation analysis. 15 According to Ramachandran and Kessides (2005), Argentina’s government default o n its bonds essentially destroyed the second pillar. 16. Promise (2005). 17. See Appendix D for additional detail o n this topic. CHAPTER4 THEIMPACT OF PENSION REFORMS 18. InPeru, the pension costs for c i v i l service retirees remain high. In Argentina, the PAYG system includes a multiplicity of plans with high replacement rates and l o w retirement ages. Uruguay's PAYG system faces fiscal problems, and actuarial model- ing is not being used to inform policymakers. InBolivia, the Bank overlooked critical reforms in the o l d system while the multi-pillar system was being established. Argen- tina and Uruguay also suffered sigruficant financial crises in the late 1990s, which caused their economies to contract, despite their strong fiscal profile. 19. Econometric studies continue to be inconclusive about the impact of pension re- form on saving. For example, Walker and Lefort (2002) report that a number of stud- ies have found the direct impact of pension reform on savings to be ambiguous. Samwick (1999) using data for a panel of countries found that only Chile experienced an increase in gross national savings rates after pension reform. While Riesen and Bailliu (1997) found that the impact of pension reform on personal saving was eight times greater for emerging market economies than for advanced economies based o n data for 11countries from 1982-93, aside from Chile, the report did not include any countries assisted by the World Bank. Davis (2005) concludes that a rise in saving i s not a guaranteed outcome of a pension reform. 20. See Appendix D for a n explanation of the interaction between funded pensions, savings, and fiscal policy. 21. Econometric studies have also tried to assess these impacts. For example, Catalan et al. (2000) used econometric analysis to look at the impact of contractual savings in- struments o n capital market development, but their study also included countries with provident funds or voluntary pension systems and not countries in which the World Bank supported the development of multi-pillar systems. Other researchers have assessed a more complicated set of indicators to determine the impact of pen- sion reform o n capital markets. Walker and Lefort (2002) found that transaction costs decreased in Chile as a side effect of pension reform, but not necessarily in Peru and Argentina, possibly due to the short duration of the time series data. When they looked at seven Latin American countries, they found n o sigruficant effects of pension funds o n dividend yields, although they did increase price to bond ratios. They also did not find any significant effect of pension fund reforms o n stock market volatility. 22. By way of comparison, this i s below the 85 percent share of equities relative to GDP in Chile, or even the 35 percent share in Brazil, where funded pensions are vol- untary (based o n data for 2000). 23. The term "participation" as used in t h i s instance represents current contributors, as opposed to all workers who have ever contributed. 24. This is also noted by Ramachandran and Kessides (2005). 25. The case study evaluations do not assess the development outcome of the reform but, rather, the Bank's assistance in promoting a satisfactory development outcome. Thus, the case studies take into consideration the multitude of uncontrolled factors that may affect the reform during the reform process. The assessment of the impact of World Bank assistance followed methodology for project evaluation. It judged the outcomes, institutional development impact, and the sustainability of the results of the assistance. The outcome rating was derived as a result of three subcriteria: rele- vance, efficacy, and efficiency. 51 5. Building Institutional Capacity 5.1 One of the p r i m a r y objectives of the Bank's pension strategy i s the sound implementation o f policy reforms. Inl i n e with this objec- tive, World Bank operations have supported institutional capacity building. Chapter 5 addresses the Bank's activities for: ( 1)the admini- stration o f PAYG systems, (2) the development of actuarial capacity, a n d (3) the regulation and supervision of funded plans.1 5.2 Out o f the 68 countries that have received loans or credits for pension r e f o r m activities, including general analytic support a n d ac- tual r e f o r m measures, 52 received loans or credits specifically for in- stitutional strengthening to improve the operation and regulation of their pension systems. One h u n d r e d twenty n i n e loans or credits sup- ported institutional capacity building, and 48 percent of those opera- tions were adjustment loans or credits and 52 percent were invest- m e n t or technical assistant loans or credits (Table 7).2 Table 7. LAC Heads the Regions in Credits and Loans Supporting Institutional Capacity Building Per Capita Total I Investment and technical assistance 1 Adiustment Number Amount Number Amount Number Amount of Per of Per of Per Region loans capita loans capita loans capita AFR 14 $103.2 $1.35 9 $13.4 $0.76 5 $89.9 $2.42 EAP 5 $440.7 $1.99 2 $10.0 $0.01 3 $430.7 $3.32 ECA 59 $1,260.4 $1.81 30 $204.4 $2.06 29 $1,056.0 $15 5 LAC 40 $1,469.7 $2.14 22 $74.6 $0.26 18 $1,395.1 $4.44 MNA 5 $45.3 $1.22 1 $9.4 $0.32 4 $35.8 $1.45 SAR 6 $22.9 $0.98 3 $4.8 $0.01 3 $18.2 $0.16 Total 129 $3,342.3 1.77 67 $316.6 $1.11 62 $3.025.7 $2.49 Source: analysis of internal Bank data 5.3 The institutional development impact of the Bank's assistance to improve the actual operation of public pension systems has not been sufficient, and i t s success has been mixed. Implementation of ac- tuarial capacity has been inadequate, with some exceptions, and countries have been unable to track fund balances, creating the poten- tial for serious macroeconomic distortions. Regulatory technical assis- 53 CHAPTER5 BUILDING INSTITUTIONAL CAPACITY tance for funded pillars should b e stronger, particularly inv i e w of continuing concerns about the appropriateness o f investment guide- lines, the lack o f competition among pension providers, and high management fees and administrative costs. Improving the Administration of PAYG Pension Systems A wide range, 5.4 Support for PAYG administration has ranged f r o m financial but small audits to the complete overhaul o f contribution collection, pension amount of, payment systems, a n d client services. ECA needed substantial opera- investment tional changes at the outset of the transition, as f e w computers were loans were done available and employers k e p t workers’ records without a central client for PAYG database. LAC countries also h a d problems with their PAYG admini- administration stration, facing misreporting of entitlements and having ”ghost” pen- sioners o n the books. Twenty countries received assistance for PAYG systems; with a number of t h e m receiving multiple loans. ECA coun- tries were the most frequent borrowers, receiving more than half o f a l l loans to improve P A Y G administration. Ingeneral, loans to support pension fund administration were relatively small. The success of 5.5 PAYG Administration in Europe and Central Asia. The Bank’s investment success with information technology coordination a n d implementa- loans in ECA’s tion has been mixed. The institutional development impact of the PAYG systems Bank’s assistance in Bulgaria was highly satisfactory, with the social have been security organization increasing revenue collection and improving mixed client service. A 1993 investment loan also i m p r o v e d the management o f the Russian pension system through computerization and organ- izational improvements. 5.6 By contrast, poor implementation o f a Hungarianproject ad- versely affected the operation of its pension reform. Information tech- nology implementation in Latvia was inefficient and delayed, although the Bank’s other activities i m p r o v e d policy planningand evaluation capacity. The Moldova project for P A Y G reform experienced sigruficant problems coordinating the implementation of information technology and the redesign of business practices. An ongoing Romania invest- m e n t project has h a d implementation difficulties according to supervi- sion reports. A smaller project for the Kyrgyz Republic h a d only a modest impact on institutional development i npart because the infor- mation technology system w a s poorly implemented. Investment 5.7 PAYG Administration in Latin America. With the exception of a loans in LAC’S l o a n to Argentina, institutional support for P A Y G administration in PAYG systems LAC was not effective. The Argentinean project was successful i n have mostly achieving fiscal savings by re-registering pensioners a n d r e m o v i n g been ineffective from the rolls deceased beneficiaries and children over the age of eli- gibility for survivors’ benefits. By contrast, the Bank missed opportu- nities for improvements i nother countries. For example, Bolivia 54 CHAPTER5 BUILDINGINSTITUTIONAL CAPACITY needed comprehensive assistance t o improve i t s poorly managed state, military, police, and judicial pension systems, but this was n o t provided. The Bank also failed to encourage the Bolivian government to develop ties with the complementary occupational pension system to stop fraudulent pension payments and reduce tens of thousands of unprocessed claims. I nPeru, the Bank did not address problems in the social security agency, w h i c h lacked the capacity to administer the system or manage its reserves. In Uruguay, the Bank ignored needed development of local offices, in part, because i t did n o t have contact with officials in the Ministry of Labor and Social Security. 5.8 PAYG Administration in Other Countries. Africa received more Investment support for P A Y G administration than MNA, EAP, or SAR. Credits i n loans in PAYG Africa were provided to nine countries, where the Bank funded small systems in other pension subcomponents within m u c h larger loans? Cape Verde re- regions have ceived the majority of funding with $5.9 m i l l i o n in three projects. The been modest Bank also supported Pakistan and India with P A Y G implementation through small pension components in larger projects. China i s the only East Asian country to receive Bank assistance for P A Y G administration. The loan, amounting an estimated $20 million, was critical in launching information technology improvements i nthe pension system, although direct support for administrative reform was modest. Actuarial Forecasting 5.9 While S h a k g y does n o t provide detailed instructions on the Actuarial evaluative a n d analytical needs of governments any more than i t analysis is stipulates financial sector preconditions, actuarial analysis i s a main- essential for a spring of any government’s ability to manage a pension system. Per- government to spective reflects that concern by indicating that financial, accounting, manage its actuarial, and governance audits are essential to pension programs in pension system increasing transparency, a n d therefore accountability. Institutionaliz- ing actuarial capacity requires the establishment of a government ac- tuarial office as a permanent p a r t o f pension administration. Actuarial projections are generally made on a scheduled basis in developed market economies under a set of demographic a n d economic assump- tions, and findings are publicly available. 5.10 To this end, the W o r l d Bank has p r o v i d e d w o r l d w i d e actuarial The Bank has training using its long-term forecasting model, PROST (see Box 4). promoted The Bank has been instrumental in educating pension experts and actuarial policymakers o n the underlying determinants of pension systems, in- training c l u d i n g the impact of demographics and economic conditions. The worldwide Bank has been less successful i nhelping countries establish their own through PROST actuarial offices and PROST has not been easily adapted to fit country circumstances.4 In addition, the Financial Sector has recently started working with the Fund for the Promotion of Scientific and Technical 55 CHAPTER5 BUILDING CAPACITY INSTITUTIONAL Research (FIRST) and U S A I D t o develop a n international p r o g r a m on building actuarial capacity. Box 4. Pension Reform Options Simulation Toolkit (PROST) The W o r l d Bank‘s PROST i s a flexible, computer-based toolkit that advises researchers and policymakers o n different options for pension reform. The m o d e l is intended to be easily adaptable t o a w i d e range of country circum- stances. PROST assesses the fiscal sustainability of different pension schemes by taking into account pension contributions, entitlements, system revenues, and system expenditures over a n extended period. PROST has been used i n more than 80 countries, and as of December 2002,364 people h a d taken a one-week training program for PROST. 5.11 Forecasting in Europe and Central Asia. Bulgaria, FYR Macedonia, Latvia, Kazakhstan, and Moldova have received Bank assistance in de- veloping policy offices and actuarial units. The Bulgarian and FYR M a - cedonian actuarial units b o t h participated in the design of their multi- pillar reforms. The Latvian Ministry o f Welfare uses a macro- simulation m o d e l developed with the assistance of the Bank to forecast the financial balance of the system. InHungary, the M i n i s t r y of Finance uses PROST to monitor the financial course of the pension reform, but the projections are not published and outside access to the information i s limited. Other Central and Eastern European countries, including Croatia and Romania, do not have actuarial offices. 5.12 Actuarial modeling has been stronger i nCentral a n d Eastern European countries than in the Commonwealth of Independent States. InKazakhstan, the analysis group in the M i n i s t r y of Labor a n d Social Protection uses PROST for demographic projections, but not for long-run financial analysis. PROST i s also used in Moldova, but a n ac- tuarial office has yet to b e established, a n d W o r l d Bank staff members assist government experts whenever policy simulations are needed. In the Kyrgyz Republic, in-house capacity to p e r f o r m long-term analysis gradually f e l l by the wayside. Actuarial projections are not con- ducted, a n d there was little ownership of PROST. Both Russia and Ukraine developed actuarial models on their own because they f o u n d that PROST did not suit their circumstances. LAC’S actuarial 5.13 Forecasting in Latin America. LAC countries have less in-house capacity is actuarial capacity than those in ECA. Bolivia, the only country with weaker than functional in-house capacity, made serious mistakes in calculating the ECA’s fiscal deficit resulting from the systemic pension reform, w h i c h the Bank did not catch or correct. InMexico, the government found PROST unable to m o d e l i t s lifetime switching o p t i o n and, as a result, the Mexican government i s not able to use PROST to forecast cash flows accurately.5 Neither Argentina nor U r u g u a y has the capacity to 56 CHAPTER5 INSTITUTIONAL BUILDING CAPACITY make the systematic actuarial evaluations needed to develop options to regain fiscal sustainability. 5.14 Forecasting in Other Countries.The Bank has h a d somewhat The Bank has greater success i n attempts t o institutionalize actuarial modeling in been more China and Korea. Based on Bank recommendations, Korea conducts successful in regular actuarial reviews o f a l l public pension schemes every five actuarial years to p r o v i d e a basis for determining subsequent parametric assistance in changes. In China, one o f the k e y aspects of World Bank assistance EAP has been PROST training for national and regional experts; the trans- fer of knowledge a n d experience was substantial. Yet data problems and the fragmentation of the pension system make it difficult t o insti- tute a national office. 5.15 The Bank has provided l i m i t e d assistance for policy evaluation The Bank has in some countries in Africa through small investment loans, but t h i s not provided has not l e d to the creation of actuarial offices. While pension audits much actuarial a n d actuarial assessments should constitute the first steps i n design- assistance in ing pension reforms, this did not occur before Cape Verde6 and Sene- AFR g a l both took actions to design a proposed multi-pillar system. Improving the Regulation of Funded Pension Systems 5.16 Funded systems have been criticized for high administrative Bank efforts a n d marketing costs, lack of competition a m o n g pension providers, have yet to a n d restrictive investment guidelines. Although costly marketing was address the recognized as a deficiency o f the Chilean r e f o r m relatively early on, high costs of no Bank investment loans or credits have specifically focused on de- funded systems veloping options to reduce costs. N i n e loans in LAC and eight in ECA supported capacity enhancement for the implementation of f u n d e d pillars. Senegal and Cape Verde also b o r r o w e d to build capacity for f u n d e d pillar implementation. Inaddition, the Bank i s working with OECD to set up regulatory templates, as w e l l as a n international pen- sion regulator-and-supervisor coordinating body. 5.17 Regulation in Europe and Central Asia. The success of Bank as- Progress in sistance in strengthening second pillar regulation varied considerably. strengthening The governments of Kazakhstan and Russia have not always been ECA's convinced o f the need for international technical assistance and have regulatory either preferred to rely on their internal expertise or on national con- framework has sultants. InKazakhstan, technical assistance was p r o v i d e d through a been mixed reallocation of funds f r o m a n earlier investment l o a n to support the pension reform, but only a small proportion of the reallocated funds were spent. D u e to the l i m i t e d disbursement, the direct objectives o f the loan were unmet.7 However, the project completion report notes that if success i s measured by improvements i npension regulation a n d administration directly attributable to Bank project supervision, the non-disbursing loan actually achieved many o f i t s development 57 CHAPTER5 BUILDING CAPACITY INSTITUTIONAL objectives. Later evaluations of Russian technical assistance for multi- pillar regulation were unfavorable, as the 1998 financial crisis delayed the adoption of a proposed pension law, and eventually the Bank halted the r e f o r m because it was not w e l l prepared.8 5.18 InCroatia and FYR Macedonia, regulatory assistance was nei- ther efficient nor timely. The initial intent of the Croatia loan was to help establish the supervisory system for the mandatory f u n d e d pillar, but Croatia m o v e d ahead with its pension reform before the Bank could finish project preparation. A s a result, the regulatory structure i s more costly than it w o u l d have been with stronger Bank support. Two small FYR Macedonian projects included subcomponents to establish the regulatory structure of the funded system. The second pillar opera- tional environment i s more complex and expensive than necessary. Furthermore, the incomplete organization of the regulatory agency was one of the factors that delayed the start of the funded pillar.9 In LAC, more 5.19 Regulation in Latin America. Bank loans for second-pillar regu- needs to be lation have been quite small, although support to LAC through ad- done to justment loans has been substantial. Loans for the supervision and liberalize regulation of funded pensions often were guided by n a r r o w terms of investment reference related to specific government requests. In Argentina, sub- regulations stantial improvements were made in accountability a n d the le- gal/regulatory system. InMexico, the government s t i l l needs to liber- alize pension fund investment regulations and institute proper regulation o f voluntary occupational pensions. (But with Mexico’s considerable in-country expertise, regulatory shortcomings are not a result of inadequate technical assistance.) Bolivia a n d C h i l e also re- ceived Bank support for the regulation of funded pensions. The larg- est loan for second pillar-regulation was to Nicaragua, although the multi-pillar r e f o r m i s now on hold. 5.20 Serious problems are found in the regulatory structure of three other LAC countries -Colombia, Costa Rica, and the D o m i n i c a n Re- public. While Costa Rica and the Dominican Republic have active loans, they do not address funded pillar regulatory issues. Although Colombian regulatory officials are aware of industry problems, n e w n asset investment guidelines are needed t o permit greater flexibility i nCosta Rica, steps should b e taken to reduce the dominant allocation. I market share of public commercial banks, and bank-related pension funds should not b e automatically assigned to undecided workers. Summary and Conclusions 5.21 The Bank has y e t to provide sufficient assistance inbuilding institutional capacity in pension administration, a n d actuarial evalua- naddition, continuing assistance i s needed to i m p r o v e second- tion. I pillar regulation. These shortcomings are in p a r t a result of incom- 58 CHAPTER5 INSTITUTIONAL BUILDING CAPACITY plete needs assessments. The failure o f administrative projects ap- pears to b e related to inadequate W o r l d Bank and client supervision, particularly in countries in w h i c h capacity i s the weakest. 5.22 PROST is a n invaluable tool for W o r l d Bank analysis of pension system design, for teaching and training client countries about factors that affect pension system balance, and for simulating a variety of re- forms to ensure a financially sustainable system. However, better di- rected technical assistance i s required to ensure that clients develop lo- cal expertise to create and maintain their own models, assess the fiscal balance of ongoing pension programs, and use actuarial modeling to design policy reforms. Many countries trying to implement PROST have been unsuccessful due to lack of training. (Also, training require- ments to create in-house actuarial expertise vary substantially by coun- try.) Some clients f o u n d PROST data requirements too intensive, or needed greater flexibility to create country-specific applications.‘o 5.23 Bank loans to establish regulatory systems for funded pensions have been l i m i t e d innumber and scope. M o r e successful regulatory and supervisory structures i nHungary and Poland were supported by policy dialogue rather than investment projects or technical assistance. T h i s may also have been the case inLAC. InLAC, investment limita- tions and unregulated voluntary pensions create a substantial future risk for participants. Future assistance in implementation will b e needed i nRegions where Bank activities have been limited, including Africa, the M i d d l e East, and Asia, providing an opportunity for the Bank to b e more proactive. NOTES 1.The chapter’s findings o n institutional development impact are based o n case stud- ies, ICR Reviews, and interviews with Bank staff. 2. Of those, 19 borrowed only for investments or technical assistance, while another 6 borrowed only for adjustment support. 3. The countries are Cameroon, Cape Verde, Ghana, Mali, Mauritius, Mozambique, Niger, Senegal, and Tanzania. 4. Even fewer countries have received formal support from the Bank to institute of- fices for strategic planning, operational planning, policy development, program monitoring, and policy evaluation. 5. The Mexican government requires annual actuarial reviews of all its pension systems, which are carried out by Mexican actuarial firms that use internationally accredited actuaries. But without in-house actuarial capacity, the independence and accuracy of these calculations can be politically compromised. 6. There i s one licensed actuary in Cape Verde who received PROST training and is fully involved i nthe reform work. 7. The limited disbursement was due in large part to the government’s subsequent decision not to borrow for technical assistance. 8. The OED Russia case study also raised substantive issues about the Russian govern- ment’s current reform, which was developed without Bank assistance. 59 CHAPTER5 BUILDING CAPACITY INSTITUTIONAL 9. Also see Chapter 3. 10. In addition, clients need another budgeting tool other than PROST to develop short-term budget estimates and minor benefit adjustments. 60 6. World Bank Coordination 6.1 The Bank's internal a n d external relationships affect the out- comes o f Bank activities a n d the success of the reforms. This chapter investigates the Bank's coordination o n pension reform activities, spe- cifically with: (1)Bank units and teams, (2) other donors a n d interna- tional organizations, and (3) client countries. 6.2 The Bank continues to struggle to establish a n informed deci- sion-making process, partly due to a lack of internal Bank coordina- tion. Because pension r e f o r m encompasses a number o f disciplines, the current sector-based resource allocation does not facilitate funding for potential pension projects. Thus, the Bank has n o t always allocated i t s resources consistently a n d efficiently in accordance with the need for a n d client interest in reform. Inaddition, the Bank lacks detailed guidelines for the design of country-specific pension strategies. 6.3 Cooperation with other international agencies and bilateral donors is also a factor in establishing stronger pension reforms. Coor- dination with other international organizations i s impeded by differ- ences in perspective on pension r e f o r m and implementation. I naddi- tion, w h i l e the Bank has w o r k e d successfully with many governments, i t needs to work harder to gain the support of a l l the ministries and stakeholders i n v o l v e d in pension reform. Coordination among World Bank Groups 6.4 Internal Bank coordination i s important given the multi-sector A lack of nature of pension reform, w h i c h affects fiscal policy, the financial sec- coordination tor, a n d social protection of the population. The Bank's matrix man- among Bank agement requires carefully planned coordination. Unfortunately, lack sectors has led of internal cooperation sometimes slows Bank assistance. Inaddition, to inconsistent pension r e f o r m i s a n ongoing process that needs constant monitoring. country policy Insome situations, Bank assistance lacked such necessary continuity on a regional and/or country-specific basis.1 6.5 A large number of sector boards have prepared and supervised loans and credits that included pension r e f o r m components (Table 8). On the one hand, social protection has taken the lead by inaugurating the Pension Primer series (a compendium of Bank-commissioned pa- pers on pension reform issues), developing the PROST model, and providingpension experts to work on operations inmany Regions 61 CHAPTER 6 WORLD BANKCOORDINATION where sector expertise was n o t available. On the other hand, other sec- tors also have strong interests i npension reform, particularly i nmacro- economic and financial sector issues.2 Differences i nperspective across nthe preparation and supervi- units have l e d t o inconsistent strategies i sion of a number of country operations.3 However, no sector has h a d a monopoly on operational effectiveness, as sector management and country outcomes are statistically uncorrelated.4 Table 8. Economic Policy and Social Protection Sectors Had a Greater Role in Pension Reform Activities Than Other Sectors Sector 1 AFR EAP ECA LAC MNA SAR 1 Total NUMBER OF PROJECTS Economic policy Financial I 1 5 10 1 3 34 6 10 14 4 1 4 0 11 58 34 Public sector governance 2 0 14 13 2 5 36 Social protectiona 0 1 30 9 1 1 42 Other 1 9 2 9 11 1 2 1 34 Total 1 26 7 93 57 9 12 1 204 ALLOCATED TO THE PENSION COMPONENT (MILLIONS us$) AMOUNT Economic policy $78.4 $200.00 $477.1 $278.10 $32.2 $22.4 $1,088.2 Financial $10.6 $302.72 $32.6 $433.8 $4.9 $0.0 $784.6 1 ~ ~~ Public sector governance 1 5 9 $0.0 $168.2 $409.3 $4.5 $19.6 $607.6 Social protection $0.0 $5.0 $808.5 $1,913.0 $25.0 $9.4 $2,760.9 Other $27.3 $10.7 $51.7 $33.2 $9.4 $7.5 $139.9 Total 1 $122.2 $518.4 $1,538.1 $3,067.5 $76.0 $59.0 I $5,381.2 Source: Pension database. a. The share of pension projects managed by Social Protection increased after the Bank's 1997 reorganization. The Bank also 6.6 Inconsistency i n the Bank's pension assistance can also b e at- lacks consistent tributed to the lack of specific guidelines on how a n d w h e n to support guidelines on pension reform. A s a result, Bank country assistance afforded too little how and when support to some countries, a n d too m u c h to others.5 In addition, turn- to support over in regional Bank leadership can exacerbate inconsistency a n d pension reform lack of continuity, especially as Country Assistance Strategy priorities change. Further, w h e n conflicts arise between the Sector a n d Country Units, there i s no agreed-upon m e t h o d of resolution.6 6.7 Even w h e n regional sector units have been interested in coordi- nating pension activities, funding has not always been forthcoming. For 62 CHAPTER 6 WORLDBANKCOORDINATION example, inAfrica, implementation of a regional pension reform pro- gram has been difficult to achieve. Moreover, w h e n Bank clients ask for Pension reform assistance, funding is n o t always available. For instance, due to the programs face high cost of the World Bank's major conferences, those w h o are able to the challenge of attend do n o t necessarily have the greatest interest or need to l e a r n limited funding about pension reform. The Bank has been addressing this issue, how- ever, through t h e greater use of distance learning. 6.8 Another disconnect in Bank coordination has been between Restricted assessments conducted by the Financial Sector Advisory Program access to (FSAP) and pension reform projects. InFYR Macedonia, although the FSAPs serves Financial System Stability Assessment (FSSA) indicated that the pre- as an additional n place, the Bank was already assist- conditions f o r reform were n o t i obstacle to ing FYR Macedonia in pension reform. Inmany countries, access t o good policy FSAP documents has been l i m i t e d due t o confidentiality require- formulation ments. W h i l e country team access to FSAP findings has improved, this has been primarily due t o i n d i v i d u a l agreements rather than a formal, collaborative Bank policy. Cooperation with Other Donors and international Organizations 6.9 The W o r l d Bank regularly collaborates with multilateral insti- Effective tutions and bilateral donors o n pension r e f o r m regionally and cooperation with w o r l d w i d e (Table 9). Effective cooperation with other international outside donors agencies and bilateral donors has resulted in stronger pension re- has resulted in forms, particularly by extending grant funding t o countries that were stronger pension unwilling t o use W o r l d Bank funding for technical assistance. reform Table 9. Many Organizations Work with the World Bank on Pension Reform Key bilateral US. Agency for International Development (USAID), Department for partners International Development (DFID),a Japanese Trust Fundb, Swedish International Development Agency (SIDA), Asian Development Bank (ADB), Inter-American Development Bank (IDB), German, Dutch, Danish, and Japanese governments. Key multilateral International Monetary Fund (IMF), International Labor Organization (ILO), partners International Social Security Association (ISSA), United Nations Development Program (UNDP), European Union (EU) Note: EU assistance has also been through PHARE and TAClS programs, a, Formerly the KnowHow Fund. b, Through the PHRD Grant. 6.10 Unfortunately, it i s n o t always easy t o ensure coordination, as many donors have predetermined w o r k programs.7 F e w discussions n the field with the European Commission o n pen- have taken place i sion reform, despite the accession of eight Central a n d Eastern Euro- pean countries, and basic agreement with the Bank on the direction of 63 CHAPTER6 WORLD BANK COORDINATION Cooperation reform. Similarly, relations with the OECD have been l i m i t e d to for- among donors m a l meetings? Cooperation i s also hindered by differences i nper- is often spective on pension reform a n d implementation. hindered by factors outside 6.11 The Bank a n d the IMF have h a d extensive ongoing discussions the Bank's on pension issues, particularly concentrating on the fiscal framework. purview And there has been successful collaboration o n assessments by the FSAP. The Bank and the IMF have failed t o reach consensus on reve- nECA, however, where the Bank's inability to n u e collection issues i fund a collaborative study has been an obstacle to better coordina- tion.9 On the whole, however, the Bank-Fund relationship is satisfac- tory. 6.12 The Bank's relationships with the Inter-American Develop- m e n t Bank (IDB) and Asian Development Bank (ADB) have a history of independent activity intermingled with collaboration. The IDB and ADB have been as likely to work separately as together on the same countries. At the start of the 199Os, the IDB deferred to the Bank, w h i c h h a d a comparative technical advantage on pension r e f o r m is- sues. Since then, the IDB developed greater financial expertise and has become more independent. Until recently, the ADB approached pension r e f o r m primarily from a financial perspective, but a m o r e re- cent ADB report from its independent evaluation department sug- gests ADB i s revisiting its underlying assumptions for social policy.10 Bank relations 6.13 The Bank has h a d successful and unsuccessful experiences with other working with USAID, ranging from disagreements about basic strate- donors has n strategy were gies to well-coordinated collaboration. Differences i been complex, evident inMontenegro and Ukraine at specific times. Yet i nBulgaria with some clear and Kazakhstan, U S A I D collaboration was instrumental inensuring successes and that those reforms were effective. The most successful Bank-USAID col- limited laboration has been for reforms supported by adjustment lending on coordination in the p a r t of the World Bank and implemented by U S A I D consultants. other instances Relationships with Clients 6.14 The Bank's relationships with clients have varied across pro- jects and countries. In some cases, such as FYR Macedonia, Kazakh- stan, Mexico, and Peru, 1 1 the Bank supported government policies without addressing their deficiencies sufficiently w h e n these policies In some cases, deviated from best practice. A s a result, this acquiescence m a y have the Bank did not compromised the long-term goals of pension policy i n these countries. communicate effectively with 6.15 In other situations, the Bank either did not communicate w e l l country with i t s clients or failed to gain the support of a l l of the government. governments For example, i n U r u g u a y a n d Hungary, the Bank h a d a g o o d relation- ship with the M i n i s t r y of Economy/Budget a n d Planning Office a n d M i n i s t r y o f Finance, respectively, but was unable to influence the 64 CHAPTER 6 WORLDBANKCOORDINATION country’s social security agency.12 InThailand, Bank communication with both government and other donors13 was unsuccessful, a n d a P A Y G r e f o r m was instituted despite the Bank’s objections. 6.16 Although it m a y b e advantageous for the Bank to maintain a The Bank needs dialogue with client countries in the absence of a Bank loan or credit, to be more the Bank also needs to respond appropriately to signals given by the receptive to governments. Inthe Philippines, the Bank should have lowered the signals given by intensity of the dialogue due t o government disinterest in effectively governments pursuing pension issues. Instead, the Bank tried to prepare a l o a n for ten years without success. Similarly, loan preparation activities lasted seven years i nSlovakia before a technical assistance a n d capacity building l o a n was signed with the Ministry o f Labor, but o n l y after the M i n i s t r y of Finance finally committed itself to b o r r o w i n g from the Bank. I nthese cases, the Bank lacked flexibility, and spent considera- bly time unsuccessfully trying to negotiate with a reluctant client. 6.17 Aside from the government, other institutions m a y also in- The Bank needs fluence pension policy. The Bank needs to address the concerns of a l l to balance the stakeholders as w e l l as the interests of the government to formulate concerns of all effective policy and assess the level o f support for i t s policies, w h i c h stakeholders can b e difficult to gauge, especially in a democracy. For example, in involved in some countries, such as H u n g a r y and Poland, independent social se- pension reform curity institutions h a d little desire to implement f u n d e d systems, w h i l e in other countries in ECA a n d LAC, private financial institu- tions, including asset managers and insurance companies, have h a d a vested interest in promoting funded pensions. Some of these institu- tions have the power to influence political decisions, and the Bank needs to better manage such diverse interests and take t h e m into ac- count m o r e effectively w h e n allocating resources. 6.18 Lastly, exogenous economic and demographic factors affect Exogenous and the outcome of a country’s pension reform. Some of these have l e d to demographic the expansion of Bank assistance. For instance, the Asian crisis factors affect prompted Korea to request Bank assistance, w h i c h helped strengthen pension reform the pension system. By contrast, other factors have reversed progress on pension reform. For example, economic crises, as in Russia and Argentina, have slowed systemic reform, weakened financial markets, a n d left private pension portfolios less diversified d u e to a higher concentration i n government debt. Conversely, the oil boom in Ka- zakhstan eased the fiscal position, a l l o w i n g the government to e n d prematurely discussions o n outstanding pension issues.14 I nsub- Saharan Africa, the HIV epidemic altered the demographic structure, creating pressure on the elderly to care for an ever-expanding number of orphaned grandchildren, with important ramifications for poverty among the aged (Kakwani and Subbarao 2005). 65 CHAPTER6 WORLD BANKCOORDINATION Case Study Evaluation of World Bank Performance Bank 6.19 As pension r e f o r m i s an ongoing process with long-term bene- performance is fits, the effectiveness of the Bank’s performance-as opposed to de- related to the velopment outcome-needs to b e considered over time. The case consistency of studies evaluated the full portfolio of bank efforts, from AAA to its approach Loans and Credits, a n d found that Bank performance varied w i d e l y and the depth of across countries (Table 10). Three factors made the m o s t important its analysis contributions to a n unsatisfactory r a t i n g of Bank performance: (1)in- consistency in the Bank’s approach, (2) lack of attention to a particular issue, such as coverage, and (3) insufficient analysis. By contrast, case studies inw h i c h the Bank’s performance was rated highly satisfactory stressed (1)good sequencing of assistance, (2) consistency with the country’s conditions, and (3) good analysis. Table 10. What Do Case Studies Show About World Bank Performance? Summary of Country Quality at entry Supervision performance China Satisfactory Highly satisfactory Highly satisfactory Bulgaria Highly satisfactory Highly satisfactory Highly satisfactory Korea Satisfactory Highly satisfactory Satisfactory Latvia Highly satisfactory Satisfactory Satisfactory FYR Macedonia Satisfactory Satisfactory Satisfactory Kazakhstan Satisfactory Satisfactory Satisfactory Hungary Satisfactory Satisfactory Satisfactory Albania Satisfactory Satisfactory Satisfactory Brazil Satisfactory Highly satisfactory Satisfactory Mexico Satisfactory Satisfactory Satisfactory Argentina Unsatisfactory Satisfactory Satisfactory Russia Unsatisfactory Unsatisfactory Unsatisfactory Peru Unsatisfactory Satisfactory Unsatisfactory Uruguay Unsatisfactory Satisfactory Unsatisfactory Bolivia Unsatisfactory Unsatisfactory Unsatisfactory Kyrgyz Republic Unsatisfactory Unsatisfactory Unsatisfactory Source: Case Study Reports. Note: The indicators cited in this table were specified in the terms of reference for the consultants conducting the case studies, as Annex 1 in the approach paper ( 2004). Summary and Conclusions 6.20 The World Bank’s pension r e f o r m activities have lacked con- sistency for several reasons. First, the Bank has not p r o v i d e d detailed guidelines to assess the p r i o r i t y of and need for multi-pillar pension reform. Second, the Bank could take greater care in allocating re- sources according to client interest in pension reform. Third, the sec- 66 CHAPTER6 WORLDBANKCOORDINATION tor-based resource allocation for Bank activities has l e d t o cross-sector rivalries t o secure access t o budgetary resources for the development of pension projects that m a y have resulted in a lack of balance inh o w the Bank's assistance i s structured. These problems are exacerbated by staff turnover, inconsistencies across Bank Networks and shifting government priorities. Insum, the lack of coordination within the Bank has prevented it f r o m establishing a transparent decision- making process. 6.21 The Bank can strengthen its pension reform activities by more frequent and substantial coordination with other international agencies and bilateral donors. Despite improvements i ncooperation, there are still a number of unresolved issues o n pension reform. The Bank would also find it easier to engage countries t o implement pension reforms by working with a broader group of ministries and considering the posi- nthe country's reform process. tions of a l l stakeholders involved i 1 .This has posed a problem in Africa and the Middle East, and also i nBolivia, Korea, NOTES the Kyrgyz Republic, and Uruguay. For Korea, its earlier graduation from the IBRD nthe Bank's dialogue. was a proximate cause of the lack of continuity i 2. After the 1997 Bank reorganization, organizational units involved i npension re- form issues included the Development Economics vice presidency, country units, sec- tor units, and networks. Similar structural units took part i na dynamic pension dia- logue before the reorganization. 3. For example, intersector disagreements have been of concern i npension operations inBosnia & Herzegovina, Georgia, India, and Ukraine. InUkraine, it was a lengthy process before the World Bank could come to any agreement o n the direction of re- form, and even now, the current reform may be too early. 4. The analysis was based o n the number of operations or the total amount of loan funds under management. 5. For instance, Mauritius, which had met the preconditions for a multi-pillar pension reform, received too little funding and Zambia, where performance was poor, re- ceived a sigruficant amount for pension reform. 6. A problem that affected FYR Macedonia, Nicaragua, and Nigeria 7. These include TACIS, UNDP, and USAID. 8. However, the OECD has had a number of staffing overlaps with the Bank, resulting nan informal transfer of knowledge. i 9. The IMF considers it more efficient for tax administrations to collect social insur- ance contributions. By contrast, the Bank found that social security offices inany number of cases were better prepared in practice, if not in theory, to collect contribu- tions. 10. See Technical Assistance Performance Audit Report on the Reform of Pension and Provi- dent Funds in Selected Developing Member Countries, Asian Development Bank. 2003. TPA: REG 2003-31. 11.However, Kazakhstan cancelled the last tranche of a pension adjustment loan be- cause there was little government ownership of the later conditions, and Kazakhstan no longer needed adjustment lending. 67 6 CHAPTER WORLDBANKCOORDINATION 12. Other international organizations, particularly labor ministries and social security administrations, occasionally provide different interpretations than the Bank o n Bank interactions with the government. 13. These include ADB, IMF, and USAID. 14. This includes coverage, annuity provision, and women’s pensions. 68 7. Findings and Recommendations 7.1 The Bank's strategy to implement multi-pillar pension reforms i s evidenced by a legacy of operational work, ESW, training, a n d seminars. Reforms have differed regionally and by country as a result of client concerns a n d Bank experience. Inother words, the Bank has followed a n approach that has differed according to country condi- tions and has n o t implemented a "one size fits all" strategy as some critics have maintained. 7.2 Nonetheless, the Bank's advice has n o t always been effective or consistent. This final chapter reviews the findings of the evaluation a n d provides recommendations for management to improve Bank ef- fectiveness. recommendations include: ( 1 )additional research on out- standing issues, (2) enhanced development of client capacity, (3) a m o r e structured approach to policy design; and (4) i m p r o v e d internal a n d external coordination. This strategy also requires a cost-effective w a y to identify key concerns to reap the rewards o f greater effective- ness. Findings 7.3 The Bank's focus on pension r e f o r m m o s t often has been sparked by concerns about fiscal sustainability, particularly w h e n mismanaged systems have created demands on the government's budget that have crowded out other expenditures and/or l e d to fiscal deficits and subsequent macroeconomic instability. The focus of client country interest in pension r e f o r m often has also been o n fiscal con- straints. And these concerns are extremely important as a pension sys- t e m that i s fiscally unsustainable will hinder growth a n d f a i l to meet i t s commitment to the aged. Nonetheless, w h i l e addressing funding gaps, too often the Bank has not addressed sufficiently the p r i m a r y goal of a pension system to reduce poverty a n d provide adequate re- tirement income within afiscal constraint. Moreover, i t has also focused insufficient attention o n the income of the aged. 7.4 nm a n y Despite this shortcoming, pension r e f o r m activities i Central and Eastern European countries have i m p r o v e d the potential for long-term fiscal sustainability overall, w h i l e providingadequate retirement income. Many reforms i n LAC have also i m p r o v e d fiscal sustainability. The Bank's activities have encouraged private pension 69 CHAPTER7 AND RECOMMENDATIONS FINDINGS plans to develop participant choice between funded systems and PAYG systems, participant choice among pension funds, and coher- ent regulatory structures to prevent f r a u d a n d abuse. ESW has cov- ered a w i d e range of topics, with particular success in the area of fis- cal analysis, actuarial modeling, and regulatory structures. Quite a f e w countries have h a d i m p r o v e d PAYG administration, i n c l u d i n g the implementation of actuarial offices. 7.5 The Bank has also emphasized the pro-growth aspects of multi- pillar reform-that i s increased savings and capital market develop- ment. But the evaluation found f e w countries in w h i c h these promised outcomes have been achieved. Currently, there i s insufficient analysis to determine the extent t o w h i c h t h i s lack of progress i s related to coun- terproductive fiscal policy or to ambiguous expected outcomes.* 7.6 Lastly, over the years, the focus o f the Bank's concerns about pension r e f o r m has evolved, f r o m supporting Chilean-type systems in L a t i n America to n e w PAYG models such as NDCs. M o r e recent ESW, for example i n Africa, has investigated the situation of the aged within the context of poverty overall. The Bank has taken strides in terms of outreach to facilitate cooperation with other international or- ganizations, although opportunities for a greater consultative process remain. 7.7 Gaps in Pension Refomzs. Based on the case studies, some o f the multi-pillar reforms supported by the Bank have shortcomings indi- cating the need for continued follow-up to the initial r e f o r m (Table 11 ) . For example, as a consequence of incomplete analysis, the Bank's n LAC tended to b e l i m i t e d to f u n d e d reforms, even w h e n activities i pensions covered a small percentage of the population, thus, as a re- sult, inadequately addressing the income of the aged. W h i l e Bank as- Table 11. Even the Best Reforms Have Continuing Challenges ~~ Post-Reform Issue Countries Low coverage Argentina (declining), Kazakhstan, Kyrgyz Republic, Peru, Russia (declining) Lack of poverty alleviation Bulgaria (women), China, Mexico, Russia, Uruguay Continuing fiscal deficits Argentina, Bolivia, Brazil, Korea (long term), Uruguay Limited actuarial capacity Kyrgyz Republic, Mexico (better modeling needed), Uruguay Underdeveloped financial sector Bulgaria, China, FYR Macedonia, Russia, Uruguay High commissions Hungary, Peru Additional pension systems Mexico, Peru High benefits Brazil, Peru Administrative assistance untimely Hungary, Latvia and ineffective Source: Case Studies, 70 7 CHAPTER AND RECOMMENDATIONS FINDINGS sistance was instrumental in instituting parametric PAYG reforms, the Bank did not press for additional first-pillar reforms required by m a n y LAC countries, such as those stemming from fragmented pen- sion systems inMexico and Peru. 7.8 The Bank persistently encouraged countries such as Ukraine a n d Russia to institute multi-pillar reforms even w h e n financial-sector conditions were weak. And the Bank failed to try to dissuade coun- tries with l i t t l e control of corruption, including Nicaragua, Russia, a n d Ukraine, f r o m actively developing multi-pillar reforms. In gen- eral, the Bank did n o t persuade multi-pillar reformers t o develop di- versified pension portfolios or support countries building the capacity to monitor the fiscal stability of their reforms. Last but not least, the Bank's performance i ni m p r o v i n g contribution collection in countries l i k e H u n g a r y and Latvia was ineffective. 7.9 The Relationship between ESWand Operations. While it i s unclear how prior ESW has l e d to adequate policies, the general focus of Bank ESW has influenced the issues considered in Bank operations. The prominence of Bank support of multi-pillar systems is the m o s t strik- ing example of h o w the Bank's strategy l e d to the preparation of pen- sion projects. 7.10 While the overall approach to support multi-pillar reform has been clear, ESW is lacking o n some specific issues and research and policy analysis has been incomplete or sporadic. For example, income of the aged has not been a priority research area or a priority for pen- sion reform. Similarly, greater analysis i s still needed on a number of financial and regulatory issues, including basic research on how t o im- prove capital markets i ncountries with multi-pillar systems and the ex- tent t o w h i c h societal corruption hinders regulation. 7.11 ImpZementing CZient Capacity. Inmany instances the Bank has not included sufficient capacity building in i t s i n i t i a l agenda or i n later follow-up activities on pension reform. In some cases, technical assistance has been successfully tied to an adjustment operation-but not always. Given client reluctance to b o r r o w for technical assistance, n e w initiatives are needed, including effective donor coordination. Within the Bank, the World Bank Treasury Department's Pension As- set Advisory Service i s a promising initiative.* Another innovation i s the collaboration of the WBI with particular client countries, particu- larly i f the clients are w e l l chosen and the lessons can b e expanded t o assist others. 7.12 Internal and External Cooperation. Internal coordination has n o t been consistent or sufficient in many areas, i n c l u d i n g advice on the income of the aged and financial sector assessment (FSAP included). Externally, the World Bank has l i m i t e d its dialogue to clients or gov- 71 CHAPTER7 AND RECOMMENDATIONS FINDINGS ernment departments that shared the Bank’s views on pension re- form. Coordination with other donors and agencies has not always been smooth. Recommendations 7.13 Based o n these findings, this evaluation has the following recommendations: DEVELOP TO DESIGN GUIDELINES PENSION REFORMS AND PAY GREATER ATTENTION TO PARAMETRIC REFORMS a. Pay greater attention to parametric reforms to ensure fiscal sustainabil- ity and t o the macroeconomic, financial, and institutional sector pre- conditions necessay for a multi-pillar reform. This would involve pre- paring and implementing guidelines t o ensure well-tailored assistance to country conditions a n d consistent policy prescrip- tions including statistical indicators and in-depth assessments. b. Be more realistic i n presenting the benefits ofthe secondary objectives of n dialogue with client countries, as there i s insuffi- pension refom i cient empirical evidence to support the claims that f u n d e d systems have or can improve savings and capital market development. BUILDCLIENT CAPACITY c. Develop a checklist for client capacity requirements (including contri- bution collection, contributor database development, actuarial and policy analysis, regulation of multi-pillar operations) to assess cli- ent requirements a n d determine how best they can b e met. This zuould involve ensuring that a plan for technical assistance i s put i n place for reform initiatives so that client capacity i s developed. CONDUCT RESEARCH ON OUTSTANDING ISSUES d. Ensure that adequate analysis i s conducted on key issues such as in- come o f the aged, the impact of corruption a n d governance o n the feasibility of effective pension regulation, methods to foster com- petition a m o n g pension funds, guidelines for investment alloca- tion, the design of non-contributory systems, a n d w a y s i nw h i c h capital markets develop, as w e l l as research offering cross-country evidence on these topics. IMPROVE INTERNAL AND EXTERNAL COORDINATION e. Develop a process to ensure that cross-sector issues are considered in- c l u d i n g financial issues identified by the FSAP and m a i n t a i n closer coordination between the Development Economics vice presidency, the Networks, sector units, and country units. 72 CHAPTER7 AND RECOMMENDATIONS FINDINGS f. Develop a strategy to play a greater role in consensus building among stakeholders, in particular, other international organizations a n d client agencies. 1.In addition, there remains considerable controversy among economists about t h e NOTES impact of pensions o n saving and the impact of saving o n growth. 2. The initiative would share expertise w i t h pension funds through seminars and technical assistance in conjunction with portfolio management. The Treasury already provides this type of assistance to central banks. 73 Appendix A Views on Pension Reform: A Brief Literature Survey 1. Over t h e years, a substantial b o d y of literature o n pension policy a n d pension r e f o r m has been developed focusing o n two fundamental issues: ( 1)the appropriate m i x of earn- ings-related pensions and poverty-reduction benefits (termed the Bismark/Beveridge con- troversy after its best k n o w proponents, see Box AA), a n d (2) the role o f funding. The con- cept o f funding i s quite o l d a n d historically has been applied i nmany different ways i n various countries. 2. Most w o r k o n social security has been o n the earnings-related p o r t i o n o f old-age provision. T h e first economic theory of social security can be traced back to Samuelson’s (1958) article that the equilibrium rate of return t o P A Y G pension plans equals the rate of population growth, under constant real wages. Aaron (1966) completed this insight by showing that i n a mature P A Y G plan, the real return equals population g r o w t h plus the rate o f g r o w t h of productivity (real wages). Buchanan (1968), Friedman (1972), a n d B r o w n i n g (1973) advocated switching t o a funded system (even before the Chilean reform) and main- taining social security commitments by issuing government bonds. Later, Ferrara (1982) a n d Weaver (1981) advocated a gradual phase-out of the PAYG system i n the U n i t e d States. Other economists have continued to find sufficient justification for traditional social security old-age benefits including Pechman, Aaron, and Tausig (1968), and D i a m o n d (1977). 3. This debate intensified with the ”money’s worth” controversy, with Feldstein’s (1974) empirical finding that the U.S. Social Security system h a d a negative impact o n sav- ing. Leimer a n d Lesnoy (1982) contested this conclusion, showing that a p r o g r a m m i n g error influenced Feldstein’s outcome. Barro (1974) argued against Feldstein’s hypothesis on theo- retical grounds suggesting that savings were n o t reduced but were shifted to bequests. Feld- nnumerous other papers. stein continued t o support h i s research i 4. Empirical evidence o n the savings controversy has been inconclusive a t best, al- n a n intense dialogue about the impact o f substituting f u n d e d sys- though i t has resulted i t e m for PAYG plans, relying o n the positive effect of such a substitution on economic growth. Other economists have h a d a narrower focus, estimating the impact of different pension provisions o n labor supply a n d the capital market, including the impact o f para- metric changes in the retirement age and the tax rate. Studies of the i m p a c t of voluntary employer-sponsored pensions o n wage/pension tradeoffs and labor force participation were also pursued, although the empirical findings were ambiguous (Fields a n d Mitchell 1984; Gustman and Steimeier 1986). Kotlikoff‘s (1988) w o r k on intergenerational equity also influenced the debate about the proper structure of a pension system. With the implementa- 75 Appendix A Views on Pension Reform: A Brief Literature Suwey tion of the Chilean funded reform, US., European, a n d L a t i n American economists began to assess i t s success, and, later, that o f other L a t i n American reforms (Bosworth, Dornbusch, Laban 1994). 5. Averting p r o v i d e d an international perspective to this body o f research with i t s con- clusion that u n d e r the right conditions, a three-pillar system was optimal. Upon publication of Averting, reviews i neconomics journals commented on the volume. Turner (1995) said ”Of the many recommendations in the book, the two most controversial are i t s advocacy of a mandatory Chilean-style funded i n d i v i d u a l account system (the proposed second pillar) a n d i t s rejection of the traditional PAYG defined benefit social security system that has been the b u l w a r k of retirement income systems i nmost OECD countries.” Disney (1995) indi- cated that Averting ”never clearly states why high savings rates are important” and suggests a reduction i ncapital stock m a y b e called for i n aging populations. Nonetheless, h e accepts the broad thrust of the Bank’s policy agenda. Beattie a n d McGillivray (1995) took issue with the report’s assertion that public pension systems failed socially and economically, identify- ing shortcomings that can apply t o both public and privatized systems. 6. Between 1994 and 2001, a wealth of articles appeared on a l l aspects of pension reform. Implicit and explicit criticisms of the World Bank approaches have come from researchers such as Bosworth and Burtless (1998), and Arnold, Graetz, and Munnell(l998). Feldstein’s (1998) edited volume on privatization was generally supportive, w h i l e Bodie, Mitchell, and Turner (1995) present a cross-section of views. Gillion, Turner, Bailey, and Latulippe (2000) cover many of the topics included i nAverting for the International Labor Organization, but suggest that more options for reform are available than the ones included i nAverting. 7. Through numerous articles and books, many experts have entered into the pension debate, both influencing the Bank’s w o r k and b e i n g influenced by it. M o s t observers agree that multi-pillar reforms are appropriate i n some instances, but quite a f e w disagree with the Bank’s prescription in specific situations a t particular times. Critics of the Bank‘s ap- proach, w h i c h considers multi-pillar systems to b e best practice, include D i a m o n d and Or- szag (2002) and Barr (2000). By contrast, Feldstein (1998) a n d Schieber and Shoven (1999) tend to support the Bank’s strategy. Kotlikoff (1994) believes that pension plans should in- na fully diversified international portfolio, and for that reason i s critical of the Bank. vest i Insum, there i s no unanimity about the time to implementation a multi-pillar reform, the proper features of a multi-pillar system, or the sufficiency of parametric reforms i nmain- taining fiscally sustainability. 76 Appendix A Views on Pension Reform: A Brief Literature Survey 3ox A.1. A Four-Country Briefing of Developments Influencing Pension Policy Worldwide Modern pension policy began with the p l a n insti- In Great Britain, Sir W i l l i a m H e n r y Beveridge x t e d by the government of German Chancellor produced a report in 1942 proposing a p r o g r a m 3tto v o n Bismarck to help workers and forestall for social insurance that w o u l d provide universal h e program of the socialist movement. The 1889 pensions based o n flat contributions a n d p r o v i d e law established a pension for a l l workers i n flat benefits as a minimum standard o f living, re- trade, industry, and agriculture f r o m the age o f placing the former means-tested system for the 70 years. Subsequently, i n 1913 the pension age elderly age 70 and over. The pension system was was reduced t o 65 years. The Bismarckian made part of the National Insurance Scheme i n scheme was based on employer and employee 1948, with a non-means-tested, basic state pen- zontributions as w e l l as o n capitalization. A state sion p a i d o u t of current revenues. Subsequently, subsidy was added to provide low-paid employ- national earnings-related programs were imple- 2es a higher pension than their contributions mented for higher-wage workers, as the flat rate warranted. After W o r l d War 1 1, PAYG financing pension was regarded as too l o w a percentage o f replaced the German-style capitalization i nm a n y earnings. F r o m this, Great Britain instituted a n social security schemes. Some modern deriva- earnings-related contributory State Earnings Re- tions of the German programs include occupa- lated Pension Scheme (SERPS) i n 1970 f r o m tional funds which are funded through book re- which employers could contract o u t i f they h a d a serves o n employer balance sheets, rather than plan providing minimum benefits. Finally, re- being directly invested i nfinancial assets. forms i n 1986 allowed individuals to contract out f r o m S E R B and establish i n d i v i d u a l accounts Inthe United States, the Social Security program (personal pensions). Initially a 2 percent govern- enacted i nthe Bis- n1935 i s earnings-related, i ment match was used t o encourage participation marckian tradition. Originally, the scheme was to nthe n e w system. In addition, there is a highly i be based o n capitalization. However, amendments developed system of occupational funds. in1939 added a number of benefits and changed the scheme to a PAYG system with only a mini- In1924, Chile became the first L a t i n American mum reserve. Insurance company executives h a d country t o adopt a social security program. By the expressed concern that the accumulation of a large time of the Pinochet government, the P A Y G sys- reserve could adversely affect the capital market, tem was i n shambles. The 1980 reform, known as encourage demands for increased benefits, and the AFP System, was one o f the many changes i n necessitate the reduction of other federal taxes. Af- nthe mid 1970s. Chile, in a process initiated i ter W o r l d War 11, voluntary employer-sponsored Chile's pension reform completely replaced the plans flourished, eventually supplemented by in- social security system with personal pension ac- dividual retirement savings options. I n1981, the counts that require pre-funded, mandatory contri- President's Commission o n pension policy rec- butions and private fund management. The n e w ommended a 3 percent contribution to mandatory pension system gave covered workers the right t o universal pension accounts based o n financial as- choose between different pension providers a n d sets invested i nindividual accounts. The 2001 between different forms of payout after their re- Commission t o Strengthen Social Security has also tirement. supported a number of options including manda- tory individual retirement accounts. 77 Appendix B Pension Systems in World Bank-Assisted Countries 1. This appendix reviews the diverse types of pension reforms that the World Bank has supported. It provides information o n the amount of Bank lending p r o v i d e d to countries by Region and type of reform, including multi-pillar a n d PAYG. 2. Nearly three-quarters of the loans were issued to ECA and LAC, the only Regions to have enacted multi-pillar systems. M o s t ECA countries maintained a relatively substantial P A Y G pillar, with pensions related to contributions, w h i l e LAC countries did not. Overall, Bank assistance for P A Y G reforms was aimed at achieving fiscal sustainability. Reforms in ECA were more l i k e l y than those i n LAC to include PAYG a n d funded-pillar assistance, as w e l l as assistance t o i m p r o v i n g social assistance. Countries with Multi-pillar Systems 3. Latin America and the Caribbean.The Bank issued loans and credits to 15 countries in L a t i n America and supported second-pillar reforms i n1 1 ,including Chile (Table B.l).1 M o s t of the Bank’s funds were lent to s i x countries: Argentina, Bolivia, Colombia, Mexico, Peru, and Uruguay.2 The Bank also provided technical assistance to multi-pillar reforms inCosta Rica, the Dominican Republic, and Nicaragua, but funding for these countries was relatively small. 4. LAC pension reforms vary considerably, although the Chilean example h a d a sub- stantial influence. Chilean consultants often p r o v i d e d advice a n d counsel, but differences from the Chilean r e f o r m abounded. Some countries included a first pillar, others allowed for choice between a PAYG system a n d a f u n d e d second pillar. Some permitted participants to switch between the two. L i k e the Chilean system, Peru’s pension system lacks a PAYG pil- lar. Peru also retains a separate c i v i l service pension system. Uruguay’s n e w system consists of a PAYG pillar, based on notional accounts similar to an NDC scheme. The f u n d e d pillar operates simultaneously with the P A Y G system and i s compulsory for contributors under age 40 whose incomes exceed a minimum. The add-on concept, i nw h i c h the f u n d e d pillar i s mandatory only for higher earners and voluntary for l o w e r earners,3 i s u n i q u e to Uruguay. 5. Bolivia’s pension system has a flat universal benefit and a f u n d e d second pillar. However, the universal pension, the Bonosol, i s not tax-financed; rather, it isfirnded u s i n g the government’s shares i n 10 privatized enterprises, a n innovation u n i q u e to Bolivia. The pension funds for the fully f u n d e d second pillar also manage the Bonosol assets. 78 Appendix B Pension Systems in World Ban k-Assisted Countries 6. Argentina, Colombia, and M e x - Table B.1 Mexico, Argentina, and Peru Received the a ico instituted pension systems that give Most Assistance Among LAC Countries Implementing a participants the choice of a PAYG sys- Mandatory Funded Pillar t e m or a f u n d e d pillar. Argentina’s sys- Amount of t e m includes a flat benefit, as w e l l as Year of loans hnberof supplemental P A Y G and f u n d e d tiers. Country reform ($ mm) loans Nearly h a l f of the provinces m a i n t a i n Mexico 199511996 $604.3 4 their own pension plans. Workers c a n Argentina 199311994 $481.I 8 make a one-time s w i t c h from the Peru 1992 $363.6 6 PAYG to the f u n d e d tiers at any time. Uruguay 1996 $149.3 5 Mexico’s reformed system i s a fully Colombia 1995 $63.7 3 f u n d e d plan, with a minimum guaran- Bolivia 1993-2000 $13.6 6 tee a n d subsidies for low-income par- ticipants. Mid-career workers can Ecuador 2001 $10.0 1 choose between the old and n e w sys- Nicaragua 2004 $8.3 tems,4 and have a lifetime switch op- Costa Rica 2000 $6.7 3 tion a t the time of retirement. InCo- Dominican lombia, a m i x e d system provides a Republic 2001 $1.5 1 choice between a defined benefit (DB), Chile 198011981 $1.4 1 partially f u n d e d P A Y G scheme and a Total $1,703.6 40 privately managed, f u n d e d plan. Low- Source:. Pension Database income contributors are subsidized and a minimum-pension guarantee i s pro- v i d e d for participants with at least 10 years of contributions i f their accounts provide a pen- sion that i s l o w e r than the guarantee. 7. Europe and Central Asia. Eleven out of 24 Bank-supported ECA countries imple- mented multi-pillar reforms, including Bulgaria, Hungary, Kazakhstan, and Latvia (Table B.2). The Bank assisted ECA countries i n improvingfiscal stability a n d reducing demo- graphic pressures by strengthening the relationship between contributions and benefits, most often through funding. Relative to LAC, reforms were more l i k e l y to b e phased inby age cohort and are only mandatory for younger workers. Older participants choose between the existing or old PAYG system and the f u n d e d tier, but their resulting placement i n the funded tier i s usually binding. 8. Multi-pillar systems implemented i n Bulgaria, Croatia, Hungary, and Latvia, include a fairly substantial contribution-based P A Y G pillar. Hungary’s f u n d e d pillar i s mandatory for n e w entrants and voluntary for those already covered by the P A Y G plan. Latvia has an NDC PAYG pillar with a guaranteed minimum pension, as w e l l as a f u n d e d tier. Partici- pants aged 50 and older remained i n the old system, workers under age 30 h a d to join the f u n d e d system, and those between 30 and 49 years c o u l d choose between the two. Similarly, Croatia’s system retained participants over age 50 i n the PAYG system, placed those under age 40 i nbetween the option to choose their sys- n the n e w f u n d e d scheme, a n d gave those i tem. Bulgaria’s second pillar was mandatory for those younger than age 40. 9. Kazakhstan i s the only ECA country that implemented a Chilean-style reform, w h i c h transferred a l l participants to a fully f u n d e d planwith a minimum-pension guarantee for those meeting contribution requirements. However, u n l i k e Chile, past obligations are fi- 79 Appendix B Pension Systems in World Bank-Assisted Countries nanced on a current basis rather than Table 8.2. Kazakhstan, Russia, Ukraine, and Hungary through recognition bonds. Inaddition, Received the Greatest Amount of Assistance Among contributors are g i v e n a choice between ECA Countries with Mandatory Funded Pillars a private and a state-run plan managed by government-chosen asset managers. Year of Amount of Number of Country reform loans ($ mm) loans Kazakhstan 1998 $323.8 4 Countries with PAYG Systems Russia 2002 $287.8 6 Ukraine 2004 $147.0 5 10. Latin America and the Caribbean. Hungary 1998 $124.1 4 Although the majority of Bank loans i n Romania 2004 $58.7 6 LAC supported a combination of first a n d second pillar reforms, the Bank Croatia I998~002 $52.1 3 also supported p u r e P A Y G reforms in Bulgaria 2000 $47.3 4 four countries (Table B.3). First pillars Lithuania 2002/2004 $26.5 2 supported by Bank operations ranged FYR f r o m p u r e P A Y G system t o partially Macedonia 2000/2002 $26.2 a f u n d e d systems or systems with no- Slovak tional accounts. Republic 2004 $25.4 2 Latvia 1995 plus $20.9 4 11. A m o n g the p u r e P A Y G reform- Poland i998 $2.6 2 ers, Brazil, the largest recipient of Bank Total $1.I 15.8 48 assistance in LAC, instituted a p l a n re- Source: Pension Database. sembling an NDC plan. Bank assistance was focused on reducing inequities between public a n d private sector workers, u p g r a d i n g municipal management, strengthening regulatory regimes for private sector workers, reduc- ing pension pressures o n the fiscal deficit, a n d increasing pension benefits for r u r a l and dis- abled workers. Jamaica's r e f o r m included grants to the elderly. 12. M a n y parametric reforms among LAC countries with multi-pillar systems focused on strengthening the safety n e t by centralizing disparate pension regimes and raising con- n Colombia, and improvingcollections and administration, such as i tributions, such as i n Argentina a n d Colombia. Argentina targeted non-contributory pensions to the poor a n d eliminating arrears. The Bank's projects supported the reduction of contribution rates in Ar- gentina, Mexico, and Peru. 13. Europe and Central Asia. The Bank supported a large number of small loans for para- metric reforms i n ECA (Table B.4). The Table B.3, Brazil Received the Majority of Bank majority o f Bank assistance for the first Assistance for PAYG Reforms in LAC pillar was intended to improve fiscal sustainability by lowering the depend- Amount of loans ency ratio, ensuring the timely pay- Countrv I$ mml Number of loans m e n t of pensions, a n d p a y i n g off ar- Brazil $1,326.3 9 rears. A number o f systems also Panama $24.5 3 eliminated privileged pensions to c i v i l Jamaica $8.9 2 servants.5 Honduras $4.2 2 Total $1,363.9 17 Source: Pension Database. 80 Appendix B Pension Systems in World Bank-Assisted Countries 14. The Bank increasingly sup- Table B.4. Turkey Received the Largest Loan Among ported the development of NDC plans ECA Countries Without a Mandatory Funded Pillar n ECA, in its pension r e f o r m l e n d i n g i Amount of loans inconjunction with other r e f o r m op- Country ($ mm) Number of loans tions, including DB f o r m u l a and point Turkey $197.7 3 systems. Kyrgyz Republic, Latvia, a n d Bosnia & Poland adopted NDC systems, w h i l e Herzegovina $43.5 6 FYR Macedonia, Hungary, a n d Moldova $37.8 4 M o l d o v a maintained DB formulas. Kyrgyz Republic $33.9 4 Credits in Bosnia-Herzegovina sup- Serbia $25.2 5 ported DB P A Y G systems. Bulgaria, Georgia $14.7 6 Croatia, Serbia, a n d Slovak Republic a l l adopted systems that transformed the Uzbekistan $10.0 1 pension formula from one that w a s ad- Armenia $8.9 5 justed on a n ad hoc basis to a point sys- Slovenia $7.7 1 tem, similar to that u s e d in Germany. Albania $7.1 5 Turkey’s PAYG reforms were insti- Azerbaijan $5.9 1 tuted t o stem operating losses, Tajikistan $2.9 1 strengthen the system’s organizational Turkmenistan $0.6 1 structure, a n d increase pensions. Legis- Total $422.3 45 lation to implement administrative re- Source: Pension Database. forms i s awaiting passage. 15. Reforms in ECA were more l i k e l y to include non-contributory assistance for the aged t h a n those in LAC. For example, in Latvia a n d FYR Macedonia, the Bank assessed the full social protection system rather t h a n Table 8.5. Zambia Received the Greatest Amount of just the employment-based pensions. Funding in Africa for Pension Reform Poverty benefits for the elderly are Amount of loans generally a residual p a r t of a larger so- Country ($ mm) Number of loans Cia1benefits System i nIllany transition Zambia $68.8 4 economies. The difference i s that these programs h a d i m p r o v e d design and Senegal $19.4 2 administration. Madagascar $5.4 1 Guinea-Bissau $5.0 1 16. Afvzca. The Bank p r o v i d e d small Mauritius $4.9 2 loans to a number of African countries Cape Verde $4.0 3 to h e l p stabilize and restructure c i v i l Cameroon $4.0 2 service pensions and p r o v i d e n t funds Ghana $3.9 3 (Table 8.5). The goal w a s to reduce the pension liability to the overall budget Mali $1.7 1 by strengthening the link between con- Mozambique $1.5 1 tributions a n d benefits and by review- Sierra Leone $1.2 1 ing investment policies to i m p r o v e re- Niger $0.9 1 turns. The Bank also supported Tanzania $0.8 2 expanding coverage in Cape Verde a n d Uganda $0.8 2 p a y i n g off arrears in Guinea-Bissau Total $122.2 26 and Sierra Leone. ~ Source: Pension Database. 81 Appendix B Pension Systems in World Bank-Assisted Countries 17. Other Regions. Except for two large loans t o Korea, EAP, MNA, a n d SAR received o n l y a f e w small W o r l d Bank loans for pension reform (Table B.6). InSouth Asia, credits to I n d i a were p r o v i d e d t o improve actuarial forecasting and reduce liabilities in the pension fund.6 I nMNA, the Bank has promoted contractual savings through the establishment o f voluntary f u n d e d systems. Fragmented pension systems were integrated in Morocco and the introduction of DC plans in Jordan was explored. Table B.6. Support to Korea for Pension Reform Far Exceeded That for Other Countries in Asia or MNA EAP SAR MNA Amount Amount Amount of loans Number of loans Number of loans Number Country ($ mm) of loans Country ($ mm) of loans Country ($ mm) of loans Korea $501.9 3 India $32.5 6 Morocco $34.4 2 China $10.0 2 Pakistan $17.5 3 Tunisia $29.9 2 Laos $5.7 1 Sri Lanka $9.0 2 Jordan $7.7 2 Mongolia $0.8 1 Afghanistan $0.1 1 Djibouti $2.5 1 Algeria $1.2 1 Yemen $0.3 1 Total $518.4 7 Total $59.0 12 Total $76.0 9 Source: Pension Database. NOTES 1.This review i s based o n the L A C country case studies by Rofman, San Martino, and Valdes-Prieto. 2. A l l except Colombia are represented i nthe case studies. 3. The government also provides a subsidy for lower-earners in the first pillar if they join the funded system. 4. The pre-reform PAYG system has been dismantled, but workers that choose the old system’s benefit package will have their entire accumulation paid in a lump sum to Treasury, which issues an annuity through the new system. The government then assumes partial investment risk for the individual account. 5. Management notes that there are n o longer any separate schemes for civil servants in ECA, except inTurkey. However, according to the US. Social Security Administration’s website o n the pension systems of other coun- tries, Albania and Russia s t i l l have special schemes for c i v i l servants and Azerbaijan has special schemes for cer- tain categories of civil servants. 6. A recent reform of the state c i v i l service plan inIndia creates a funded pillar that the government hopes to augment with reformed provincial plans and contributions from the self-employed. 82 Appendix C Performance Ratings for Pension Projects 1 . This appendix reviews the development outcome ratings of completed pension re- form projects, compares t h e m to other Bank projects, and describes the rationale for the rat- ing. The loans and credits have been d i v i d e d into three groups: (1)those that are 80 to 100 percent devoted t o pension reform, (2) those that are between 30 a n d 80 percent devoted to pension reform, and (3) those i nw h i c h pension issues account for less than 30 percent of the loan.1 The rationale for the division i s to investigate whether loans that p r i m a r i l y focused on pension r e f o r m performed better than those in w h i c h pension r e f o r m was only a small com- ponent of the loan. Table C.1. Performance Ratings for the Project Overall Institutional development impact Total evaluated projects Outcome satisfactory Sustainability likely substantial ~ Commit- No. of Commit- No. of Commit- No. of Commit- No. of ments projects ments projects ments projects ments Project projects ($mm) (%I (W (W (%) h r o) (%I All pension projects 140 $29,697 87% 82% 77% 75% 57% 43% Pension component >80% 10 $2,506 100% 100% 70% 68% 80% 54% Pension component 30%-80% 10 $1,859 70% 49% 80% 78% 50% 11% Pension component <30% 120 $25,332 88% 83% 78% 76% 55% 44% Total World Bank Projects 3,391 $259,968 70% 76% 56% 66% 38% 43% Notes: Two pension projects were not rated for institutionaldevelopment impact and were excluded from the calculations. Outcome satisfactory (highly satisfactory, satisfactory, moderately satisfactory). Sustainability likely (highly likely, likely). Institutionaldevelopment impact substantial (high, substantial). Pension projects from approval fiscal year 1984 to 2005. World Bank projects from approval fiscal year 1984 to 2003. 2. The development outcome of the pension component was scored by u s i n g a two- p a r t rating system (satisfactory or unsatisfactory) based on a review of ICRs,* ICR reviews a n d PPARs.3 The development outcome for the overall project i s from the ICR reviews and the PPARs. The ratings for the overall project are based o n the six-part r a t i n g scheme, 83 Appendix C Performance Ratings of Pension Projects nthe pension compo- w h i c h was condensed t o a two-part system, equivalent to that used i n e n t analysis.4 3. On average, overall ratings for projects including a pension component were higher than a l l Bank projects for outcome, sustainability, and institutional development (Table C.1). On a disaggregated basis, this finding was true for projects entirely devoted to pensions and those with a small (30 percent or less) pension component. The ratings for this m i d d l e group were largely influenced by a large l o a n to Russia that was rated moderately unsatisfactory.5 A m o n g a l l pension projects, the group of pension projects with pension r e f o r m comprising 80 to 100 percent of the loan's objectives was rated the highest in terms of development out- come (see also Table C.2).6 Table C.2. Performance Ratings for the Pension Component Total evaluated projects I Outcome satisfactory Commit- Pension Commit- Pension No. of ments component No. of ments component Project projects ($mm) ($mm) projects (Yo) (%) (W All pension projects 134 $28,262 $4,784 75% 77% 89% Pension component >80% 10 $2,506 $2,503 100% 100% 100% Pension component 30%- 80% 10 $1,859 $736 80% 50% 64% Pension component (30% 114 $23,898 $1,544 73% 77% 82% Analysis of Pension Performance Rating 4. Loans made pvimarilyfor pension reform. Pensions accounted for over 80 percent of the n 7 percent o f the loans evaluated. By Region, they accounted for 24 percent of pension loan i n ECA and 1 lending i 1percent in LAC. The loans included seven adjustment loans, two in- vestment loans, and one technical assistance project. These 10 projects a l l h a d v e r y high de- velopment outcome ratings (Table C.3).7 5. Loans issued to Argentina, FYR Macedonia, and Uruguay were a l l r a t e d highly satis- factory. The large adjustment loan made to Argentina helped consolidate Argentina's pen- sion systems and quickly l e d to sigmficant savings. A 1998 Uruguay loan i m p r o v e d the regulation and operations of the funded system's pension plans, i n particular through the diversification of assets! Another highly satisfactory pension loan was the small technical assistance loan to FYR Macedonia w h i c h successfully assisted the legal r e f o r m for the P A Y G system a n d a framework l a w for a multi-pillar system.9 6. The other loans were rated satisfactory as some aspects of l o a n performance could n Brazil w a s advanced but have been improved. The r e f o r m of the social security system i nthe 1999 loan; the conditions of the 2000 loan were fully complied with, n o t completed i 84 Appendix C Performance Ratings of Pension Projects but did not reduce special privileges. The Kazakhstan r e f o r m was successful inimproving contributions, although the third Loan tranche was not completed. The L a t v i a n l o a n w a s ex- emplary in achieving a full reform of the pension system, but the information technology component w a s unsatisfactory. The Mexican project implemented a n e w private pension system, but the complementary reforms for the housing fund p r o v e d difficult to achieve d u e nPeru, a pension reform was instituted, but the public systems con- to political constraints. I tinue t o lack adequate records or controls. Table C.3. Performance Ratings: Loans 80%-100% Devoted to Pension Reform Activities Pension component Country Loan Year Loan outcome Sustainability Gmm) Argentina Provincial Pension Reform Adj. 1997 Highly satisfactory Likely $620 Project Brazil Social Security Special Sector Adj. 1999 Satisfactory Likely $758 Loan Brazil Second Social Security Special Sector 2000 Satisfactory Likely $505 Adj. Loan Bulgaria Social Insurance Administration 1997 Satisfactory Highly likely $24 Project Kazakhstan Pension Reform Structural Adj. Loan 1998 Satisfactory Not rated $300 Project Latvia Welfare Reform Project 1997 Satisfactory Likely $15 FYR Pension Reform Tech, Asst. Project 1999 Highly satisfactory Likely $1 Macedonia Mexico Second Contractual Savings Dev. i998 Moderately Uncertain $400 Project satisfactory Peru Pension Reform Adj. Loan 1997 Satisfactory Likely $100 Uruguay Contractual Savings Structural Adj i998 Highly satisfactory Uncertain $100 Loan 7. Loans made Zargelyfor pension reform. Seven percent of a l l completed loans and credits h a d pension components that made up 30 to 80 percent of the loan. This includes 12 adjust- m e n t loans and 7 investment loans, with development outcomes for the project overall rang- ing f r o m highly satisfactory t o unsatisfactory (Table C.4). These ratings generally correlated with the rating o f the pension component. 8. The pension component of the Peruvian Financial Sector Adjustment L o a n w a s r a t e d satisfactory, as the social security reforms were expected to i m p r o v e the longer-term resil- iency of the system. The pension component of the Mexican Contractual Savings Develop- m e n t Program was also rated satisfactory, although the ICR indicated that no progress had been made i n (subsidized) housing loans or for the nthe r e f o r m of pension assets invested i r e f o r m of the public pension system. 9. By contrast, the pension component i n the Russian Social Protection Adjustment L o a n was rated unsatisfactory because the financial crisis h a d sigruficant negative effects on the prospects for adopting a n d implementing pension reform. Similarly, the 1993 H u n g a r i a n 85 Appendix C Performance Ratings of Pension Projects investment loan did n o t help support pension reform or implementation w e l l even though the r e f o r m was successful overall. Table C.4. Performance Ratings: Loans 30%-80% Devoted to Pension Reform Activities Pension Pension component component Country Loan Year Loan outcome outcome ($mm) Albania Tech. Asst. for Social Safety Net Project 1994 Satisfactory Satisfactory $3 Capital Market Dev. Tech. Asst. Argentina Project 1994 Satisfactory Satisfactory $3 Fin. Markets & Pension Reform Tech. Bolivia Asst. Project. 1996 Satisfactory Satisfactory $3 Hungary Public Sector Adj. Loan 1998 Satisfactory Satisfactory $93 Pension Admin. & Health Insurance Hungary Project 1993 Unsatisfactory Unsatisfactory $15 ’ KYrgYz Republic Social Sector Adj. Credit 1999 Satisfactory Satisfactory $26 KYrgYz Moderately Republic Social Safety Net Project 1995 unsatisfactory Satisfactory $7 Contractual Savings Development Mexico Program 1997 Highly satisfactory Satisfactory $200 Moderately Peru Second Fin. Sector Adj. Loan 1999 satisfactory Satisfactory $136 Moderately Russia Social Protection Adi. Loan 1997 unsatisfactory Unsatisfactory $249 10. Large pension component expenditures, small pension component shares. Of the 114 evalu- ated loans, 37 o f the loans a n d credits contained a pension component amounting to $10 million or more, although the median pension component share was less t h a n 10 percent of the full loan (Table C.5). The development outcome r a t i n g for the project overall often did n o t correlate with the rating for the pension component. 11. The Argentinean Special Structural Adjustment Loan was rated unsatisfactory but the pension component was satisfactory because n e w pension eligibility criteria were insti- tuted the legislative and executive branches to i m p r o v e the program’s efficiency and c u r b abuses. Similarly, w h i l e the PPAR rated the 1996 Zambian loan moderately unsatisfactory, the pension component was satisfactory because pension reforms were eventually imple- mented, even though they were delayed.10 12. By contrast, the pension components for Hungary, Kazakhstan, and U r u g u a y were deemed unsatisfactory even though the overall loans were satisfactory. According to the PPAR for the 1987 Uruguay loan, the Bank and government did not work w e l l together. In Kazakhstan, pension arrears were n o t cleared - a k e y condition the loan. Similarly, in Hun- gary the financial stability in the pension system w a s increased, also a k e y objective of the loan. Inaddition, the r e f o r m of the social safety n e t w a s brought to a standstill and imple- mentation was postponed indefinitely. 86 Appendix C Performance Ratings of Pension Projects 13. The pension components of loans to Bosnia-Herzegovina, Romania, Tunisia, a n d Ukraine also were unsatisfactory, although the overall loan was satisfactory. InTunisia, the preparation of the social security r e f o r m was delayed because studies took longer than an- ticipated a t the time the second tranche was released. Similarly, the pension component of the Romanian loan was unsatisfactory because studies addressing the long-term issue of fi- nancial viability o f the P A Y G schemes were not carried out. 14. Ln Bosnia, the earliest loan, the 1997 Transition Assistance Project, was satisfactory, but the pension component was unsatisfactory because the Bank remained unconvinced that pro- grams proposed w o u l d b e either fiscally sustainable or provide i m p r o v e d targeting. By con- trast, two subsequent 1998 and 1999 loans h a d satisfactory pension components and were sat- isfactory overall. Ukraine's 1995 loan was moderately satisfactory, w h i l e the pension component was unsatisfactory as measures h a d n o t been introduced to strengthen the social safety net, in contrast to the loan objectives. While the Bank and the government agreed on the principals of the reform by third tranche release, implementation was delayed. 15. InKorea, three very large pension components of loans undertaken i n1998 a n d 1999 were rated satisfactory. These loans started a national pension scheme that was to p a y out full pensions starting in 1998. Inaddition, a "Compensation Fund" w a s started to finance a n immediate means-tested, non-contributory social pension for the elderly. Measures were in- stituted to pave the w a y for opening up pension fund investments insecurities other t h a n government bonds or directed investments. The second structural adjustment loan l a i d groundwork for additional reforms of the pension system. Table C.5. Performance Ratings: Loans Less Than 30% Devoted to Pension Activities with a Pension Component of At Least $10 million Pension Pension component component Country Loan Year Loan outcome outcome ($mm) AFR Moderately Zambia Fiscal Sustainability Credit 2000 satisfactory Satisfactory $44 Economic Recovery & Investment Moderately Zambia Credit 1996 unsatisfactory Satisfactory $22 EAP Korea, Rep. Structural Adj. Loan Project 1998 Satisfactory Satisfactory $225 Korea, Rep. Second StructuralAdj. Loan 1999 Satisfactory Satisfactory $200 Korea, Rep. Economic Reconstruction Loan 1998 Satisfactory Satisfactory $77 ECA Bosnia- Herzegovina Public Fin. Structural Adj. Credit 1998 Satisfactory Satisfactory $11 Bosnia- Second Public Fin. Structural Adj. Herzegovina Credit 1999 Satisfactory Satisfactory $11 Bosnia- Herzegovina Transition Asst. Project 1997 Satisfactory Unsatisfactory $10 Bulgaria Social ProtectionAdj. Loan 1999 Satisfactory Satisfactory $16 Croatia Structural Adj. Loan 2002 Satisfactory Unsatisfactory $24 Moderately Hungary Second Structural Adj. Loan 1991 satisfactory Unsatisfactory $10 Fin. & Enterprise Development Kazakhstan Project 1995 Satisfactory Unsatisfactory $13 Lithuania Structural Adj. Project 1997 Satisfactory Satisfactory $13 87 Appendix C Performance Ratings of Pension Projects Pension Pension comDonent comDonent Country Loan Year Loan outcome outcome ($mm) Lithuania Second Structural Adj. Loan 2001 Satisfactory Satisfactory $13 Moderately Moldova Second Structural Adj. Loan 1998 satisfactory Satisfactory $20 Romania Structural Adj. Loan 1992 Satisfactory Unsatisfactory $17 Moderately Russia First Rehabilitation Loan 1995 satisfactory Satisfactory $20 Employment Services & Social Moderately Russia Protection Project 1993 satisfactory Satisfactory $12 Moderately Ukraine RehabilitationLoan 1995 satisfactory Unsatisfactory $83 Ukraine Coal Sector Adj. Loan 1997 Satisfactory Satisfactory $20 Ukraine Programmatic Adj. Loan 2002 Satisfactory Non-evaluable $20 Uzbekistan RehabilitationLoan 1995 Unsatisfactory Satisfactory $10 LAC Argentina Special Structural Adj. Loan 1999 Unsatisfactory Satisfactory $101 Special Repurchase Facility Argentina Support Loan 1999 Highly unsatisfactory Unsatisfactory $19 Rio de Janeiro State Reform- Brazil Privatization Project 1998 Satisfactory Satisfactory $21 Social Protection Special Sector Brazil Adj. Loan 2000 Highly satisfactory Satisfactory $16 Programmatic Fin. Sector Adj. Brazil Loan 2001 Satisfactory Satisfactory $12 Moderately Colombia Structural Fiscal Adj. Loan 2002 satisfactory Satisfactory $50 Colombia CO Programmatic FSAL I 2003 Satisfactory Satisfactory $10 Programmatic Human Moderately Ecuador Development Reform Loan 2003 satisfactory Unsatisfactory $10 Panama Economic Recovery Loan 1992 Highly satisfactory Satisfactory $18 Peru Structural Adj. Loan 1992 Highly satisfactory Satisfactory $56 Peru Fin. Sector Adj. Loan 1992 Highly satisfactory Satisfactory $42 Uruguay Second Structural Adj. Loan 1989 Satisfactory Satisfactory $12 Uruguay First Structural Adj. Loan 1987 Satisfactory Unsatisfactory $11 MNA Moderately Morocco Contractual Savings Dev. Loan 1998 satisfactory Satisfactory $25 Economic & Fin. Reforms Support Tunisia Loan 1992 Satisfactory Unsatisfactory $25 NOTES 1.These divisions were made based o n the likelihood that the strength of supervision would depend o n the level of Bank resources, that is, thorough pension supervision would be more likely when the proportion of the loan devoted to pensions was higher. 2. These are self-evaluations by Bank teams. 3. A simple satisfactory/unsatisfactory scale was used because in many ICRs where the pension component was not prominent, a more detailed verbal evaluation of the pension component was not available to make a nu- anced judgment. Inaddition, six projects were deemed "non-evaluable." 4. The six-part project-rating scheme is: highly satisfactory, satisfactory, moderately satisfactory, moderately un- satisfactory, unsatisfactory, highly unsatisfactory. The two-part equivalent scheme categorizes highly satisfac- tory, satisfactory and moderately satisfactory ratings as satisfactory; and moderately unsatisfactory, unsatisfac- tory, highly unsatisfactory ratings as unsatisfactory. 88 Appendix C Performance Ratings of Pension Projects ~~ 5. In particular, t h e outcome rating and institutional development ratings weighted by commitments are particu- larly low. The low ratings in outcome and institutional development are the result of the $800 million Russian Social Protection Adjustment Loan, which received respective ratings of moderately unsatisfactory, and modest. This project makes up 46 percent of total commitments for the 9 pension projects i nt h i s category. The rating for institutional development impact was also influenced by a modest rating for a 1997 Mexican adjustment loan. 6. Only two projects, for Latvia and FYR Macedonia, were less than 100 percent devoted to pension reform. The other components in these projects were related to social assistance. 7. The development outcome for the pension component was satisfactory for all 10 projects. Because the rating for the project overall shows more delineation among ratings, only those ratings are shown. The development outcome ratings are uncorrelated with sustainability or size of the pension component. 8. However, development outcome of these two pension reforms as they stand today were adversely affected by the collapse of the Argentinean and Uruguayan economies. 9. The loan was evaluated o n its performance in supporting an associated adjustment loan rather than for the success of the pension reform. A further assessment by FSAP indicated that FYR Macedonia’s financial sector and the regulatory structure were not ready for the pension reform, and progress o n the reform had to be slowed down. 10. However, the Queisser, Bailey and Woodall (1997) study indicated that the pension system remained defi- cient in design, financing, and administration. 89 Appendix D Multi-pillar Pension Systems, Transition Costs, and Savings 1 . Multi-pillar pension systems create a government revenue shortfall i f they divert contributions used to p a y current pensions f r o m the PAYG system to the f u n d e d pillar. This transition debt m a y be debt-financed or tax-financed. Often, a combination of the two i s used, with parametric PAYG reforms reducing budgetary expenditures so that the transi- t i o n cost i s less. Countries can also fund the transition costs of multi-pillar pension reforms by issuing recognition bonds and/or n e w government securities, w h i c h m a y b e purchased by the pension funds. ECA countries tended to have high pension liabilities before the tran- sition, so that a Chilean-style r e f o r m w o u l d have created a n extremely high transition deficit requiring sigruficant reductions i n expenditures or sigruficant increases i n government debt with resulting high fiscal deficits. For that reason, the choice of a m o r e substantial P A Y G pil- lar was wise. By contrast, LAC pension systems were m u c h smaller d u e t o restricted cover- age. Chilean-style systems were more feasible, as the transitional debt was considerably lower. 2. Debt-Financed Transition. If the reform i s a partial privatization of the pension system without reductions i nother expenditures, it will be totally debt-financed. Inthat case, the im- pact on national savings should b e roughly neutral because the revenue losses will b e fully offset by the increase i nprivate savings (the flows to the mandatory second pillar accounts). In other words, the increase i nexplicit public debt is offset by the decrease i ni m p l i c i t debt, that is, the obligations t o future pensioners from the o l d PAYG system. However, a debt-financed transition also could have a n e t negative impact o n saving if the interest rate on explicit debt is higher than the implicit interest rate on the former implicit debt (the rate of r e t u r n to the P A Y G system). Inthat case, the explicit debt m a y increase more than the i m p l i c i t debt de- clines. A second w a y i nw h i c h a debt-financed transition could have a negative impact on sav- ings is if financial markets reacted adversely t o the growth of the explicit debt, even with a de- crease i nimplicit debt. A s a result, the interest rate on the explicit debt c o u l d increase because of a perception of higher risk. A completely debt-financed transition will not have any capital market effect, and pension funds will primarily hold government bonds. 3. Tax-Financed Transition. A tax-financed transition i s the t e r m used to describe a fiscal adjustment that offsets revenue losses from the diversion o f contributions to accounts that are either partly or fully funded. Such deficit reductions can come from: ( 1 )tax increases, (2) reductions in other expenditures, (3) a parametric PAYG r e f o r m that creates a surplus offset- ting the revenue losses from the transition to funding. Ineach case, the reduction i n public savings would be smaller than the increase i n private savings, leading to an increase inna- tional savings, similar to that created by other types of fiscal adjustments. 90 Appendix D Multi-pillar Pension Systems, Transition Costs, and Savings 4. Changes in Personal Saving. Personal savings m a y react to the reform itself but these changes are empirical and uncertain. Changes i npersonal savings due to pension reform are de- termined inlarge part by the importance of borrowing constraints that exist inmost couniries. 5. nthe retirement age. According to the overlap- M o s t reforms include a n increase i ping generations model, a n increase i n private nthe retirement age should lead to a decline i savings because w h e n employees work for a longer period, they do not need to save as m u c h to achieve their optimal retirement income goals a n d reach the savings level that would smooth their consumption levels over their retirement. Therefore, private savings should decline. However, most reforms also include reductions i nbenefits through changes n indexation (from wages t o prices) or direct changes i i n the benefit formula. According to the same overlapping generations model, this should lead t o an increase i nprivate savings, for analogous reasons- the changes reduce retirement income a n d disturb o p t i m a l savings plans to smooth consumption over the life cycle. Therefore, workers need to save m o r e to restore o p t i m a l consumption smoothing. But, consider a worker who i s partly contributing to a second pillar i nw h i c h the expected returns are higher than those o f the PAYG system. This m a y reduce voluntary savings d u e t o the income effect, or increase t h e m d u e to a sub- stitution effect. M o s t economists conclude that the income effect would dominate, but the n e t impact i s l i k e l y to b e very small. 91 92 93 a Q) 94 2 95 97 99 P 3 0 E .- u) I B 5 3 s 0 ._ v) w I 101 103 E 0 rn b U 0 Q) v) > P u) .- U E B U Q) 0 E rn E F n U E Q) E 0 Q Ei 0 .- E 0 tn E a n 2 a E I I! 105 106 107 108 7 ~ 1 .. 8 t 1 d i I 1 I t iI 1 . 1 I I 109 I 110 s m '~ 2 - T x N Lo d c\i m N d m t s L 0 $J 22 B .- I ? e L 111 112 113 I 1 I 114 I I 115 116 L > P .cI E E 2 - w . v) E a E 5 E a 0 U a 1 E U 44 t c .- c: Q: E n .- S C (: I S - Q: n 7 - Y C nc C c C c .- c C d d h lL a i 3 117 b g c w B U * .- c a E n u .- S C + E e E .- T S S - a s UY c s W .- L. E d E Z " - O E 0 - 8 s 'E" 0 0 - s h ~~ s F7 t g cv 119 120 Appendix G Selected World Bank Economic and Sector Work by Region (Country-Specific) Excludes FSAP assessments ~~ ~ Report title Report type Date Number AFRICA 1 Malawi: Public expenditures - issues and options Vol. 1 Economic Report 09/30/2001 22440 2 Mauritius: Country economic memorandum : sharpening the Economic Report 04/12/1995 13215 competitive edge Vol. 1 3 The role of occupational pension funds in Mauritius Vol. 1 Policy Research Working 0413012003 WPS3033 Paper 4 The insurance industry in Mauritius Vol. 1 Policy Research Working 0413012003 WPS3034 Paper 5 Namibia s social safety net : issues and options for reform Vol. 1 Policy Research Working 10/31/1998 WPS1996 Paper 6 Senegal: Policies and strategies for accelerated growth and poverty Economic Report 04/03/2004 28143 reduction - a Country Economic Memorandum Vol. 1 of 1 7 Safety nets and income transfers in South Africa Vol. 1 Departmental Working 02/28/1999 19335 Paper 8 The use of asset swaps by institutional investors in South Africa 'I I' Policy Research Working 12/01/2003 WPS3175 Vol. 1 of 1 Paper 10 A social protection strategy for Togo Vol. 1 Working Paper (Num. 07/31/1999 20534 Series) 11 Reforming pensions in Zambia : an analysis of existing schemes Policy Research Working 01/31/1997 WPSl716 and options for reform Vol. 1 Paper EAST ASIA AND PACIFIC China: The emerging capital market Vol. 1 Sector Report 11/03/1995 14501 China: The emerging capital market Vol. 2 Sector Report 1110311995 14501 Population aging and pension systems : reform options for China Policy Research Working 05/31/1996 WPSl607 VOI. 1 Paper China: Reform of state-owned enterprises Vol. 1 Sector Report 06/21/1996 14924 China: Pension system reform Vol. 1 Sector Report OW2211996 15121 How can China provide income security for its rapidly aging Policy Research Working 10/31/1996 WPS1674 population? Vol. 1 Paper 7 Old age security : pension reform in China Vol. 1 Publication 09/30/1997 17090 8 Implicit pension debt, transition cost, options, and impact of China's Policy Research Working 02/28/2001 WPS2555 pension reform : a computable general equilibrium analysis Vol. 1 Paper 9 Reforming Indonesia ' s pension system Vol. 1 Policy Research Working 10/31/1996 WPS1677 Paper 10 Indonesia development policy report : beyond macroeconomic Sector Report 12/04/2003 27374 stability Vol. 1 of 1 121 Appendix G Selected World Bank Economic and Sector Work, by Region (Country Specific) Report title Report type Date Number 11 The Korean pension system at a crossroads Vol. 1 Sector Report 0511012000 20404 12 The national pension scheme of the Republic of Korea Vol. 1 WBI Working Paper 0113112001 22712 13 Public expenditure in Malaysia: who benefits and why Vol. 1 Publication 0113111979 10113 14 Mongolia: Poverty assessment in a transition economy Vol. 1 Sector Report 0612711996 15723 15 Financialsector reforms in Mongolia Vol. 1 WBI Working Paper 01101/ I 998 18873 16 Mongolia: Poverty Reduction Strategy Paper and Joint Staff Poverty Reduction 0811812003 26563 Assessment Vol. 1 of 1 Strategy Paper 17 Philippines: An agenda for the reform of the social security Sector Report 0912911995 13400 institutionsVol. 1 18 Philippines: Improving government performance : discipline, Sector Report 0413012003 24256 efficiency and equity in managing public resources (a public expenditure, procurement and financial management review) Vol. 1 19 Thailand: Increasing private sector participation and improving Sector Report 10111/1994 13132 efficiency in state enterprises Vol. 3 EA&ERN EUROPE AND CENTRAL 1 Albania: Beyond the crisis - a strategy for recovery and growth Vol. Economic Report 12/07/1998 18658 1 2 Household welfare, the labor market, and social programs in Publication 0513112001 WTP503 Albania Vol. 1 3 Albania: Poverty assessment Vol. 1 of 1 Economic Report 11/05/2003 26213 4 Armenia: Interim Poverty Reduction Strategy Paper and joint Poverty Reduction 0412712001 22131 assessment Vol. 1 Strategy Paper 5 Azerbaijan: Poverty assessment Vol. 1 Economic Report 0212411997 15601 6 Azerbaijan: Poverty assessment Vol. 2 Economic Report 0212411997 15601 7 Bosnia & Herzegovina: From recovery to sustainable growth Vol. 1 Publication 05/3111997 16711 8 Bosnia & Herzegovina: Public expenditure review Vol. 1 Economic Report 1112611997 17161 9 Bosnia & Herzegovina: Public expenditure review Vol. 2 Economic Report 1112611997 17161 10 Bosnia & Hetzegovina: From aid dependency to fiscal self-reliance Economic Report 1013112002 24297 a public expenditure and institutionalreview Vol. 1 11 Social safety net and the poor during the transition : the case of Policy Research Working 05/31/1995 WPS1450 Bulgaria Vol. 1 Paper 12 Managing fiscal risk in Bulgaria Vol. 1 Policy Research Working 01/3112000 WPS2282 Paper 13 Croatia: Beyond stabilizationVol. 1 Economic Report 1211911997 17261 14 Pension reform in Croatia Vol. 1 Working Paper (Num. 0212812003 25983 Series) 15 Czech Rep: Capital market review Vol. 1 Publication 0513111999 19306 16 Czech Rep: Enhancing the prospects for growth with fiscal stability Publication 0913012001 22888 Vol. 1 17 Czech pension system : challenges and reform options Vol. 1 Working Paper (Num. 0613012002 24675 Series) 18 Estonia: Public expenditure review update Vol. 1 Economic Report 0710311997 16420 19 Estonia: Country economic memorandum : implementing the EU Publication 0613011999 19404 accession agenda Vol. 1 20 Georgia: Interim poverty reduction strategy paper and joint Poverty Reduction 12/04/2000 21448 assessment Vol. 1 Strategy Paper 122 Appendix G Selected World Bank Economic and Sector Work, by Region (Country Specific) Report title Report type Date Number 21 Georgia: Poverty update Vol. 1 Economic Report 01/ I 012002 22350 22 Georgia: Public expenditure review Vol. 1 Economic Report 11/25/2002 22913 23 Hungary: Reform of social policy and expenditures Vol. 1 Publication 0413011992 10647 24 Private pension funds in Hungary : early performance and Policy Research Working oa/31/1996 wpsma regulatory issues Vol. 1 Paper 25 Poverty and social transfers in Hungary Vol. 1 Policy Research Working 05/31/1997 WPS1770 Paper 26 The Hungarian pension system in transition Vol. 1 Working Paper (Num. Series) 27 Fiscal risks and the quality of fiscal adjustment in Hungary Vol. 1 Policy Research Working 0913011999 WPS2176 Paper 28 Hungary: On the road to the European Union Vol. 1 Publication 11/30/1999 19923 29 Pension reform in Hungary : a preliminary assessment Vol. 1 Policy Research Working 07/31/2001 WPS2631 Paper 30 Kazakhstan: Living standards during the transition Vol. 1 Sector Report 0312211998 17520 31 Kazakhstan: Joint private sector assessment Vol. 1 Sector Report 0913011998 Ia467 32 Kazakhstan: An ambitious pension reform Vol. 1 Working Paper (Nurn. 01/31/2001 23156 Series) 33 Latvian pension reform Vol. 1 Working Paper (Num. Series) 34 Social transfers and social assistance - an empirical analysis using Policy Research Working Latvian household survey data Vol. 1 Paper 35 Lithuania: An opportunity for economic success Vol. 1 Publication oa/31/199a 18383 36 Lithuania: An opportunity for economic success Vol. 2 Publication oa/31/199a 18383 37 Macedonia: Focusing on the poor Vol. 1 Sector Report 06111/ I 999 19411 38 Macedonia: Focusing on the poor Vol. 2 Sector Report 06/11/1999 19411 39 Social insurance in the transition to a market economy : theoretical Policy Research Working 04/30/1996 wps15aa issues with application to Moldova Vol. 1 Paper 40 Moldova: Public expenditure review Vol. 1 Sector Report 10/09/1996 15532 41 Poland: Income support and the social safety net during the Publication 01/31/1993 11592 transition Vol. 1 42 Poverty in Poland Vol. 1 Sector Report 09/14/1994 13051 43 Poverty in Poland Vol. 2 Sector Report 0911411994 13051 44 Poland: Growth with equity policies for the 1990s Vol. 1 Economic Report 09/2a/1994 13039 45 46 ~~ Poverty and social transfers in Poland Vol. 1 ~ ~ ~~~~ ~ Understanding poverty in Poland Vol. 1 -~ ~ *- Policy Research Working Publication 03/31/ I 995 WPS1440 07/31/1995 14876 47 Wage and pension pressure on the Polish budget Vol. 1 Policy ResearchWorking 06/30/1997 WPS1793 Paper 48 Poland: Country economic memorandum : reform and growth on the Economic Report 0711511997 I6858 road to the EU Vol. 1 49 Welfare and the labor market in Poland : social policy during Publication economic transition Vol. 1 50 The quest for pension reform: Poland's security through diversity Working Paper (Num. 10/31/199a 20111 Vol. 1 Series) 123 Appendix G Selected World Bank Economic and Sector Work, by Region (Country Specific) Report title Report type Date Number 51 Shaping pension reform in Poland: security through diversity Vol. 1 Working Paper (Num. 08/31/1999 20852 Series) 52 Pension reform and public information in Poland Vol. 1 Working Paper (Num. 08/31/2000 23142 Series) 53 Disability and work in Poland Vol. 1 Working Paper (Num. 01/31/2001 23145 Series) 54 Poland: The functioning of the labor, land and financial markets: Sector Report 12/31/2001 22598 opportunities and constraints for farming sector restructuring Vol. 1 55 Romania: Poverty and social policy Vol. 1 Sector Report 0413011997 16462 56 Romania: Poverty and social policy Vol. 2 Sector Report 0413011997 16462 57 Romania: Public expenditure review Vol. 2 Economic Report 06/26/1998 17743 58 Romania: Building institutions for public expenditure management : Economic Report 08/31/2002 24756 reforms, efficiency and equity - a Public Expenditure and Institutions Review Vol. 1 59 Romania: Poverty assessment Vol. 1 of 2 Sector Report 09/30/2003 26169 60 Romania: Poverty assessment Vol. 2 of 2 Sector Report 09/30/2003 26169 61 "Poverty in Russia" in Service provision for the poor : Public and Publication 01/01/2004 28403 private sector cooperation Vol. 1 of 1 62 Income transfers and social safety net in Russia Vol. 1 Publication 09/30/1992 11168 63 The role of women in rebuilding the Russian economy Vol. 1 Publication 09/30/1993 12305 64 Russian Fed: Social protection during transition and beyond Vol. 1 Sector Report 02/02/1994 11748 65 Russian Fed: Social protection during transition and beyond Vol. 2 Sector Report 02/02/1994 11748 66 Pension funds in Central Europe and Russia : their prospects and Policy Research Working 0513111995 WPS1459 potential role in corporate governance Vol. 1 Paper 67 Russian Fed: Toward medium-term viability Vol. 1 Publication 0413011996 15559 68 "Gender Aspects of Pension Reform in Russia" in Making the Publication 12/31/1999 WDP411 transition work for women in Europe and Central Asia Vol. 1 69 Dividing the spoils - pensions, privatization, and reform in Russia's Policy Research Working 03/31/2000 WPS2292 transition Vol. 1 Paper 70 Assisting Russia's transition - an unprecedented challenge Vol. 1 publication 01/01/2002 25397 71 Russian Fed: Bank assistance for social protection Vol. 1 of 1 Working Paper 01/01/2002 27970 72 Integrating housing wealth into the social safety net : the elderly in Policy Research Working 08/31/2003 WPS3115 Moscow Vol. 1 Paper 73 Serbia & Montenegro: Medium-term public expenditure priorities Economic Report 1011612002 24880 Vol. 1 74 Slovakia: Restructuring for recovery Vol. 1 Publication 0913011994 13528 75 Slovakia: Development policy review Vol. 1 Sector Report 1113012002 25211 76 Slovakia: Development policy review Vol. 2 Sector Report 11/30/2002 25211 77 Slovakia: Joining the EU: a development policy review Vol. 1 of 1 Publication 06/0112003 26607 78 Winners and losers in transition : returns to education, experience, Policy Research Working 08/31/1994 WPSl342 and gender in Slovenia Vol. 1 Paper 79 Slovenia: Labor market issues Vol. 1 Sector Report 0313011998 17741 80 Slovenia: Economic transformation and EU accession Vol. 2 Publication 0313111999 19020 81 Turkey: Challenges for adjustment Vol. 1 Economic Report 04/01/ I 996 15076 82 Non-bank financial institutions and capital markets in Turkey Vol. 1 Publication 0413012003 25954 124 Appendix G Selected World Bank Economic and Sector Work, by Region (Country Specific) Report title Report type Date Number 83 Turkey: Country economic memorandum: towards macroeconomic Economic Report 07/28/2003 26301 stability and sustained growth Vol. 1 of 3 84 Ukraine: Reforming the pension system Val. 1 Working Paper (Num. 01/31/1996 17365 Series) 85 Pension reform, growth, and the labor market in Ukraine Vol. 1 Policy Research Working 02/28/1997 WPSl731 Paper 86 Ukraine: Public expenditure review : restructuring government Sector Report 06/25/1997 16112 expenditures Vol. 1 87 Economic growth with equity : Ukrainian perspectives Vol. 1 Publication 10/31/1999 WDP407 88 Uzbekistan: Social and structural policy review Vol. 1 Sector Report 08/25/1999 19626 LATIN AMERICAN AND THE CARIBBEAN 1 Argentina: Capital market study Vol. 1 Sector Report 12/21/1994 12963 2 Effects of social security on lifetime income distribution in Argentina Poverty & Social Policy 08/31/1995 17364 Vol. 1 Working Paper 3 Cordoba: public sector assessment : proposals for.reform Vol. 1 Sector Report 0511511996 15132 4 Cordoba: public sector assessment : proposals for reform Vol. 2 Sector Report 0511511996 15132 5 The Argentine pension reform and its relevance for Eastern Europe Policy ResearchWorking 0813111997 WPS1819 Vol. 1 Paper 6 Private pension funds in Argentina ' s new integrated pension Policy Research Working 08/31/1997 WPSl820 system Vol. 1 Paper 7 Argentina: The fiscal dimension of the convertibilityplan : a Economic Report 0112211998 16996 background report Vol. 1 8 Argentina: Financial sector review Vol. 1 Sector Report 0912811998 17864 9 The pension system in Argentina - six years after the reform Val. 1 Working Paper (Num. 0613012000 23089 Series) 10 Pension reform in Bolivia : innovative solutions to common Policy Research Working 09/01/1997 WPSl832 problems Vol. 1 Paper 11 Bolivia: Public Expenditure Review Vol, 1 Economic Report 0611411999 19232 12 Private sector and social services in Brazil : who delivers, who pays, Sector Report 0613011994 13205 who regulates Vol. 1 13 Brazil: Social insurance and private pensions Vol. 1 Sector Report 0112511995 12336 14 Effects of social security on lifetime income distribution in Brazil Vol. Poverty & Social Policy 08/31/1995 17362 1 Working Paper 15 Reforming social security : lessons from internationalexperience DepartmentalWorking 05/31/ I 997 17120 and priorities for Brazil Vol. 1 Paper 16 Labor market prospects of public employees in Brazil : an empirical DepartmentalWorking 0611711997 17069 evaluation Vol. 1 Paper 17 Brazil: From stability to growth through public employment reform Sector Report 0211711998 16793 Vol. 1 18 Brazil: From stability to growth through public employment reform Sector Report 0211711998 16793 Vol. 2 19 Brazil: Critical issues in social security Vol. 1 Publication 05/31/2001 22513 20 Broadening the base for growth : a report on the state of Bahia Vol. Economic Report 10/26/2001 21377 1 21 Rural poverty alleviation in Brazil : towards an integrated strategy Economic Report 12/27/2001 21790 Vol. 1 125 Appendix G Selected World Bank Economic and Sector Work, by Region (Country Specific) Report title Report type Date Number 22 Rural poverty alleviation in Brazil : towards an integrated strategy Economic Report 12/27/2001 21790 Vol. 2 23 Brazil: Issues in fiscal federalism Vol. 1 Economic Report 06/04/2002 22523 24 Brazil: Inequality and economic development Vol. 1 of 2 1Policy Sector Report 10/01/2003 24487 report 25 Brazil: Inequality and economic development Vol. 2 of 2 1 Sector Report 1010112003 24487 Background papers 26 The rationale and performance of personal pension plans in Chile Policy Research Working 02/29/1992 WPS867 Vol. 1 Paper 27 Chile: Pension reform and growth Vol. 1 Policy Research Working 0613011995 WPS1471 Paper 28 Chile: Social security reform and women's pensions Vol. 1 Working Paper (Num. 02/28/2001 22565 Series) 29 Chile's pension reform after twenty years Vol. 1 Working Paper (Num. 1213112001 24079 Series) 30 Gender Effects of Social Security Reform in Chile in The World Publication 0110112002 25579 Bank economic review 16 (3) Vol. 1 31 Revealed preference and self-insurance - Can we learn from the Policy Research Working 01/31/2002 WPS2754 self-employed in Chile? Vol. 1 Paper 32 Pooling, savings, and prevention - mitigating the risk of old age Policy Research Working 0513112002 WPS2849 poverty in Chile Vol. 1 Paper 33 Colombia's pension reform : fiscal and macroeconomic effects Vol. Publication 1113011995 WDP314 1 34 Colombia: Social Safety net assessment Vol. 1 Sector Report 0813012002 22255 35 Colombia: The economic foundation of peace Vol. 1 Publication 12/31/2002 25426 36 Costa Rican pension system : options for reform Vol. 1 Policy Research Working 0613011995 WPS1483 Paper 37 Costa Rica: A pension reform strategy Vol. 1 Publication 0113112000 20100 38 Dominican Rep: Poverty assessment : poverty in a high-growth Sector Report 1211712001 21306 economy 1986-2000Vol. 1 39 Dominican Rep: Poverty assessment : poverty in a high-growth Sector Report 1211712001 21306 economy 1986-2000Vol. 2 40 Mexico: Social security reform : the capital accumulation and Policy Research Working 10/31/1990 WPS512 intergenerational distribution effect Vol. 1 Paper 41 Mexico: Mobilizing savings for growth Vol. 1 Sector Report 1212311997 16373 42 Mexico: Mobilizing savings for growth Vol. 2 Sector Report 1212311997 16373 43 The 1997 pension reform in Mexico Vol. 1 Policy Research Working 0613011998 WPS1933 Paper 44 The economics of gender in Mexico : work, family, state, and market Publication 04/3012001 22242 VOl. 1 45 Mexico: Fiscal sustainability Vol. 2 Sector Report 0611312001 20236 46 Paraguay: Country economic memorandum : macroeconomic Economic Report 0310511999 18392 policies to reactivate growth Vol. 1 47 Paraguay: Financial sector review Vol. 1 Economic Report 1111512002 24249 48 Peru: Public expenditure review Vol. 1 Economic Report 1013111994 13190 49 Peru: Reforming the pension system Vol. 1 Working Paper (Num. 0613011995 17361 Series) 126 Appendix G Selected World Bank Economic and Sector Work, by Region (Country Specific) ~ ~ ~~ ~ Report title Report type Date Number 50 Pension reform and private pension funds in Peru and Colombia Policy Research Working 11/30/1997 WPS1853 Vol. 1 Paper 51 Trinidad & Tobago: Macroeconomic assessment and review of Economic Report 06/28/1996 15187 public sector reform and expenditures : the changing role of the state Vol. 1 52 Uruguay: Options for pension reform Vol. 1 Working Paper (Num. 06/30/1995 17363 Series) 53 Uruguay: Country economic memorandum Vol. 1 Economic Report 0112211996 14263 54 Fiscal impact of switching from a pay as you go to a capitalization Departmental Working 07/31/1996 17412 system : the case of Uruguay ' s largest pension system, BPS Vol. 1 Paper 55 Uruguay: Financial sector review Vol. 1 Sector Report 11/15/2000 20199 56 Uruguay: Maintaining social equity in a changing economy Vol. 1 Economic Report 07/17/2001 21262 MIDDLEEAST AFRICA AND NORTH 1 Djibouti: Pension system reform: strategic note Vol. 1 Sector Report 12/31/2001 22087 2 Egypt: Country economic memorandum : issues in sustaining Economic Report 0311511997 16207 economic growth Vol. 1 3 Egypt: Country economic memorandum : issues in sustaining Economic Report 03115/1997 16207 economic growth Vol. 2 4 Egypt: Country economic memorandum : issues in sustaining Economic Report 0311511997 16207 economic growth Vol. 3 5 The role of non-bank financial intermediaries (with particular Policy Research Working 0313111998 WPSl892 reference to Egypt) Vol. 1 Paper 6 The pension system in Iran : challenges and opportunities Vol. 1 of Sector Report 09/01/2003 25174 2 7 The pension system in Iran : challenges and opportunities Vol. 2 of Sector Report 09/01/2003 25174 2 8 Morocco: Financial sector strategy note Vol. 1 Sector Report 09/26/2000 20885 9 Morocco: Poverty update Vol. 1 Sector Report 03/30/2001 21506 10 Morocco: Poverty update Vol. 2 Sector Report 03/30/2001 21506 11 Options for pension reform in Tunisia Vol. 1 Policy Research Wo...ing 07/31/1993 WPSl15 Paper 12 Tunisia's insurance sector Vol. 1 Policy Research Working 05/31/1995 WPS1451 Paper ASIA SOUTH Subnational Administration in Afghanistan Vol. 1 of 2 / Assessment Sector Report 04/01/2004 28435 and recommendations for action Afghanistan - Subnational Administration in Afghanistan Vol. 2 of 2 / Sector Report 04/01/2004 28435 A guide to Government in Afghanistan How well do India's social service programs serve the poor? Vol. 1 Policy Research Working 08/31/1990 WPS491 Paper India: Reducing poverty, accelerating development Vol. 1 Publication 01/01/2000 20749 Maharashtra : reorienting government to facilitate growth and Sector Report 10/31/2002 25053 reduce poverty Vol. 1 Nepal: Financial sector study Vol. 1 Sector Report 10/16/2002 24959 Pakistan: Economic update : adjustment and reforms for a better Working Paper 04/22/1998 19015 future Vol. 1 127 Appendix G Selected World Bank Economic and Sector Work, by Region (Country Specific) Report title Report type Date Number 8 Pakistan: Public expenditure review : reform issues and options Vol. Economic Report 10/07/1998 18432 1 9 A framework for civil service reform in Pakistan Vol. 1 Sector Report 1211511998 18386 10 Pakistan: Reforming provincialfinances in the context of devolution Economic Report 11110/2000 21362 -an eight point agenda Vol. 1 11 Reforming Punjab ' s public finances and institutions Vol. 1 Sector Report 08/2112001 20981 12 Pakistan: Development policy review - a new dawn? Vol. 1 Sector Report 0410312002 23916 13 Household savings : an estimation for Sri Lanka Vol. 1 Working Paper (Num. 10/31/1976 SDF27 Series) 14 Promoting growth in Sri Lanka : lessons from East Asia Vol. 1 Policy Research Working 06/30/1995 WPS1478 Paper 15 Review of superannuation benefit programs in Sri Lanka Vol. 1 Sector Report 0511912000 20468 128 Attachment 1: Management Response Bank Assistance to Pension Reform and the Development of Pension Systems: An OED Review JANUARY 4,2006 129 Attachment 1 Management Response Bank Assistance to Pension Reform and the Development of Pension Systems: An OED Review’ Infroducfion 1. Public a n d private pension systems are a n essential mechanism i n client countries for reducing the risks of old-age poverty and smoothing lifetime income so the aged can m a i n - t a i n living standards. Pension systems vary substantially i nbeing sufficiently adequate to cover the risks o f old-age poverty for most of i t s population, sufficiently ufordable so as not to overburden current workers, employers a n d governments; sustuinabze to b e able to pro- v i d e promised benefits consistent with contribution levels over multiple generations and sufficiently robust to withstand the effect o f economic, political a n d demographic shocks. Many are inadequate b o t h i n terms of the level of poverty risk reduction p r o v i d e d relative to cost and the level o f coverage; m a n y create a substantial financial b u r d e n on employers, employees a n d governments in order to p a y for benefits beyond the affordability of the cur- r e n t society; many have h a d severe problems of financial sustainability creating a n obstacle to fiscal stability, economic growth and poverty reduction; and many have p r o v e n highly vulnerable to variance i neconomic and political conditions. 2. Bank Support. The World Bank i s a n acknowledged leader i n assisting countries to ensure that their pension systems are adequate, affordable, sustainable a n d robust. Since 1984, the Bank has helped 68 countries r e f o r m their pension systems with support from m o r e than 200 loans a n d credits. Moreover, the Bank has served as a central source of n e w thinking on pension reform, h a v i n g issued over 350 papers a n d publications, i n c l u d i n g books such as Averting the Old Age Crisis2 a n d Old Age Income Support in the 2 1 ~Century3 f n shaping the conceptualization of pension r e f o r m options a n d w h i c h have p r o v e n catalytic i strategies. Finally, the Bank has been a leader in setting the stage for debate and for knowl- edge gathering and management in its organization o f international, national, and regional conferences a n d seminars on pension r e f o r m issues; m u l t i p l e f o r m a l and i n f o r m a l training programs for policymakers a n d practitioners; creation and dissemination of a c o m m o n computer m o d e l for systematically projecting and evaluating pension r e f o r m options; and by creating communication linkages through i t s website, international and regional net- works. 3. The OED Review. Management welcomes this timely a n d comprehensive r e v i e w of Bank assistance to pension r e f o r m and the development of pension systems. This Manage- m e n t Response discusses the OED report’s m a i n findings and presents views o n k e y issues that are fundamental to the success of the Bank‘s work in this important area. It i s impor- tant to keep in mind that most structural reforms undertaken i nBank client countries oc- ’ Bank Assistance to Pension Reform and the Development of Pension Systems: An OED Review (CODE2005- 0087), September 23,2005. Averting the Old Age Crisis. Oxford: Oxford University, 1994. 3 O l d Age Income Support in the 2lSt Centuy: The World Bank’s Perspective on Pension Systems and Reform, World Bank, 2005. 131 Attachment 1 Management Response curred in the late 1990s w h i l e judgments over outcomes, sustainability, adequacy, afforda- bility and robustness can only b e made after a period of 10 or m o r e years. Conclusions re- garding the Bank’s assistance should therefore be v i e w e d as tentative. Main OED Findings and Recommendations 4. OED Findings. The OED reports principal findings are: Focus onfiscal sustainability. The Bank’s focus o n pension reform has often been sparked by concerns about fiscal sustainability. W h i l e addressing funding gaps, too often the Bank has not sufficiently addressed the primary goal o f a pension system to reduce poverty and provide adequate retirement income within a fiscal constraint. Moreover, it has also focused insufficient attention o n the income o f the aged. Emphasis on increasing savings and capital markets development. The Bank has em- phasized the pro-growth aspects o f multi-pillar reform-increased savings and capital market development but the OED evaluation found few countries inwhich these prom- ised outcomes have been achieved. Gaps in support to reformprograms. Some o f the multi-pillar reforms supported by the Bank have shortcomings indicating the need for continued follow-up t o the initial reform. For example, the Bank’s activities inL C R tended to be limited t o funded re- forms, even when pensions covered a small percentage o f the population. W h i l e Bank assistance was instrumental in institutingparametric pay as you go reforms, the Bank did not press for additional first-pillar reforms required by many LCR coun- . tries. Supportfor multi-pillar reforms where inadequate preconditions existed. The Bank persistently encouraged some countries t o institute multi-pillar reforms even when financial-sector conditions were weak. Furthermore the Bank failed t o try t o dis- suade countries with little control o f corruption from actively developing multi-pillar reforms. The Bank did not persuade multi-pillar reformers to develop diversified pension portfolios or support countries building the capacity t o monitor the fiscal stability o f their reforms. Last but not least, the Bank’s performance in improving . contribution collection in some countries was ineffective. Economic and sector work. While it i s unclear h o w prior economic and sector w o r k led to adequate policies, the general focus o f Bank E S W has influenced the issues considered in Bank operations. While the overall approach to support multi-pillarre- form has been clear, E S W has been lacking on some specific issues. Research andpol- i c y analysis often has been incomplete, spotty, and sporadic. Capacity building. In many instances the Bank did not include sufficient capacity building in its initial agenda or in later follow-up activities o n pension reform. In some cases, technical assistance has been successfully tied to adjustment opera- . tions-but not always. Internal and external cooperation. Internal coordination has not been consistent or sufficient in many areas, including advice o n the income o f the aged and financial 132 Attachment 1 Management Response sector assessment (FSAP included). Externally, the W o r l d Bank has limited i t s dia- logue t o clients or government departments that shared the Bank’s views o n pension reform. Coordination with other donors and agencies has n o t always been smooth. 5. O E D Recommendations. The OED Review’s recommendations are discussed below, along with Management comments. COMMENTS MANAGEMENT On the Analysis and Conclusions 6. Management finds the study comprehensive in analyzing support for pension re- forms a n d agrees with the general thrust o f most of the recommendations. Inparticular, Management agrees with the recommendations to strengthen internal coordination and the diagnostic framework for determining country readiness for privately-managed second pil- lar reforms. 7. A Dynamic Learning Framework. In Management‘s view, the r e p o r t c o u l d do m o r e to portray the dynamic character of the learning process w h i c h has characterized the Bank’s framework for pension reforms over the 20 years covered. Working at a country level a n d internationally, the Bank has contributed to knowledge products and, in the process, learned and reevaluated its position and formal and i n f o r m a l guidance to staff. This dy- namic character has influenced the development of research programs for the areas where k e y uncertainties exist a t a conceptual, empirical and implementation level. The report ac- knowledges the Bank‘s knowledge products as a foundation for policy dialogue but does n o t evaluate knowledge products as a n essential element of the country dialogue. These in- clude central policy-focused analytical products (such as the pension p r i m e r series), capac- ity building efforts (including global pension core courses, regional training programs, Bank-sponsored conferences and seminars), and policy research programs (such as on an- nuities, coverage and old-age poverty). This learning process has l e d to revisions o f the Bank’s perspective on pension r e f o r m over time, including a m u c h firmer v i e w on objectives of pension systems and reforms, a n d the substance and process criteria the Bank applies for supporting country pension r e f o r m proposals. A more systematic r e v i e w of the Bank’s knowledge products would have p r o v i d e d a richer v i e w of the w h o l e range of instruments used in helping countries reach their desired development objectives. 8. Benchmark Criteria Applied. Management would have preferred that the review apply w h a t it views as benchmark evaluation criteria appropriate to specific t i m e periods of Bank interventions. The report applies the 2001 document Social Protection Sector Strategy: From Safety Net to Springboard4 (“Strategy”) as a benchmark against w h i c h the Bank’s activi- ties from 1984 to 2004 are assessed. However, m u c h of the Bank‘s support for pension re- f o r m measures predates this document. Moreover, there was a substantial growth of learn- ing in the Bank and i nthe international community as to good practice approaches. Indeed, the report acknowledges that the ”most intense” period of Bank activity w a s from 1998- 2001, before ”Strategy” was published. M u c h of the Bank’s work during that p e r i o d was in- 4 Social Protection Sector Strategy: From Safety N e t fo Springboard, World Bank, 2001. 133 Attachment 1 Management Response fluenced by the findings of Averting the Old Age Crisis,5 which, although not a f o r m a l strat- egy document, serves as a stronger benchmark to measure the Bank's activities in the 1990s, prior to the publication of "Strategy." There are sigruficant points of difference between "Averting" and "Strategy" largely arising f r o m the experience the Bank gained through it's involvement with countries that underwent parametric and structural reforms. Manage- m e n t believes that u s i n g "Strategy" as the b e n c h a r k t o assess activities that were g u i d e d by "Averting" leads the OED Report to a more critical set of findings and conclusions than would have been the case i f more time-appropriate benchmarks were used. Finally, the re- port acknowledges the Bank guidance document Old Age Income Support in the 21s* Century: An International Perspective on Pension Systems and Reform6 but does n o t link recent interven- tions to the n e w approaches suggested in this document reflecting four years additional ex- perience, consultation, research and reflection. 9. The Impact o f the Bank. In Management's view, the report appears to overstate the role a n d potential impact of Bank support, relative to the influence o f country r e f o r m agen- das a n d the agenda of other development partners. M u c h of Bank lending took place after reforms were legislated and where countries required fiscal assistance. I nL a t i n America for example, o f the 12 countries that undertook structural reforms, the Bank was active in only seven. Of these, the Bank p r o v i d e d support against the backdrop of country initiatives often heavily influenced by the Chilean experience, with the Bank becoming i n v o l v e d only once the broad r e f o r m m o d e l h a d been decided. 10. Fiscal Necessity for Pension Reform for Growth and Poverty Alleviation. Manage- m e n t puts a greater weight than the review o n the importance of fiscal crises a n d the prior- ity this demanded as p a r t of a poverty-alleviation a n d growth strategy. The threat to eco- nomic stability and growth posed by pension systems in L a t i n America i n the 1990s arose f r o m design characteristics a n d mismanagement that posed unsustainable fiscal and eco- nomic burdens and w h i c h often benefited m i d d l e a n d u p p e r income groups a t the expense of the poor. Poorer households in L a t i n America often bore the brunt of these circumstances in inflation taxes or forgone economic growth or fiscal resources for m o r e properly targeted poverty alleviation support. InECA, many of the existing pension entitlements c o u l d not b e sustained in a n environment of substantial economic retrenchment. M a n y o f the aged poor in ECA suffered f r o m effective benefit reductions. Government requests for Bank assistance inreforming pension systems and the form this assistance took emanated from such circum- stances. I nmost cases, the Bank was approached by clients faced with crisis situations, and staff reacted with assistance often focused on stopping fiscal hemorrhaging of earnings- related public pension systems. Only once hard-won economic stability was achieved by the late 1990s could a more sober dialogue b e established about fiscally-sustainable ways of addressing more broadly the risks to income in old age (including coverage), particularly the risk of poverty. I nManagement's view, these changing client needs and Bank responses could have come o u t more clearly in the OED review. Averting the Old Age Crisis, op. cit. 6 OldAge Income Support in the2St Century: The World Bank's Perspective on Pension Systems and Reform, op. cit. 134 Attachment 1 Management Response 11. Minimum Coverage, Benefits and Poverty Alleviation. InManagement’s view, the review should have taken a broader v i e w of the links between pension r e f o r m and p o v e r t y reduction. The review for the most p a r t focuses narrowly on the l i n k s between pension re- form a n d the elderly poor. There are other essential dimensions o f the role of pension re- form i nalleviating poverty. The report questions whether the Bank’s assistance i n pension r e f o r m focused sufficiently o n supporting the elderly poor, including extending coverage a n d ensuring minimumbenefit entitlements. Two essential concepts are not discussed. It i s because relatively high income earners often have h a d substantial claims o n GDP v i a their acquired pension rights that pension r e f o r m i s a n d was urgently needed. Expenditures of such large sums o f resources on relatively f e w individuals compared t o total p o p u l a t i o n raises questions of resource allocation, equity, fairness and growth. Reducing the fiscal b u r d e n required for existing and future pension entitlements attributed to only a portion of workers a n d retirees i s an essential means of freeing up precious fiscal resources for old-age poverty alleviation. Although pension reform m a y initially only affect a f e w people, it can have strong and relevant ripple effects throughout the economy. This was p a r t of the ra- tionale for intervention i nthe pensions sector i nthe late 1990s, and Management would nthe review. have l i k e d a larger discussion of this rationale i 12. Strategy Evolution and Differentiation since 2001. The Bank has differentiated i t s strategy according to country characteristics, based on assessments of prospective vulner- ability of the elderly population i n relation to other vulnerable groups as reflected i n the 2001 Strategy. Moreover, it has weighed the fiscal resources a n d institutional capacity nec- essary to provide additional support for vulnerable populations. Attention to coverage has increased substantially i n LCR, where there i s a potential to reduce vulnerability of the aged w h i l e managing fiscal costs. InECA, where coverage has been higher than i n other regions, nthe face of weaknesses i increasing attention i s being p a i d t o coverage i n some countries in benefit provision; a n d in South Asia and Africa, expansion of coverage i s b e i n g considered against the pressing needs o f a l l vulnerable groups as w e l l as fiscal and institutional re- source constraints. The focus on vulnerability i s an essential metric of the 2001 Strategy that does not come out i nthe OED review. 13. Country Cases. Insome cases, application of n a r r o w technical criteria leads to dif- ferences between OED and Management o n interpretations of results.7 On the Recommendations Recommendation 1 . Develop Guidelines to Design Pension Reforms and Pay Greater Attention to Parametric Reforms: (a) Pay greater attention to parametric re- f o r m to ensure fiscal sustainability and to the macroeconomic, financial, and insti- tutional sector preconditions necessary for a multi-pillar reform. This would in- volve preparing a n d implementing guidelines to ensure well-tailored assistance to country conditions and consistent policy prescriptions i n c l u d i n g statistical indica- tors and in-depth assessments; and (b) Be more realistic inpresenting the benefits of the secondary objectives of pension reform i ndialogue with client countries, as there For example, the report in some cases uses the drafting o f legislation as an indication that a country has under- taken a reformprogram. Nicaragua, for example, i s cited as having undertaken a reform even though the reform was put on hold as indicated in Chapter 3 o f the report. 135 Attachment 1 Management Response i s insufficient empirical evidence to support the claims that funded systems have or can i m p r o v e savings and capital market development. 14. Parametric Reforms. Management will continue to focus attention on r e f o r m op- tions to ensure the affordability and adequacy of benefits, as w e l l as system a n d fiscal sus- tainability and robustness i nthe face of shocks. The Bank’s will continue to evaluate pen- sion systems against country objectives and then to recommend reforms measures (including parametric and/ or structural reforms) appropriate to country conditions. M a n - agement notes that the current framework used for evaluating pension r e f o r m options al- ready evaluates both parametric and structural reforms. Recently, the Bank co-sponsored with the Swedish social security administration a k e y review of a promising u n f u n d e d pen- sion r e f o r m option- Non-financial or Notional Defined Contribution (NDC) plans.* 15. A Diagnostic Framework for Second Pillar Reforms. Management fully agrees with the importance of guiding criteria for establishing privately-managed second pillar reforms a n d will establish a working group represented by HDN, FSE and PREM to develop a diag- nostic framework. The recently initiated series of FSAP Updates a n d demand-driven ESW are enabling the Bank to identify the actions needed to i m p r o v e the performance of y o u n g second pillars, as w e l l as the actions that n e w reformers need to consider before establishing second pillars. The results of this work will b e supplemented by the development o f rele- vant indicators for a larger number of reforming countries i n order to produce a meaningful set of guidelines. Attempting to standardize such criteria i n checklists would l i k e l y restrict the ability of staff to tailor advice on reforms to the w i d e variation of country conditions. 16. Impact on Savings and Capital Markets. Although the empirical evidence linking pension policies and economic growth are w i d e l y debated, strong theoretical arguments and a g r o w i n g body of empirical evidence p o i n t to a link between pension reforms a n d strengthening the efficiency and transparency of the financial sector. There are also w e l l documented strong empirical l i n k s between financial sector development a n d economic growth. To the degree that pension policy r e f o r m has a medium-term impact on growth, whether through structural reforms or parametric reforms to existing systems, such growth will b e the strongest and most efficient measure to reduce poverty, i n c l u d i n g poverty of the aged. Management agrees with the importance of realistically presenting the benefits of the secondary objectives of savings growth and capital market development. Measures to sup- p o r t these objectives m u s t b e fully consistent with a hierarchy of measures supporting a strategy towards poverty alleviation. Recommendation 2. Build Client Capacity: Develop a checklist for client capacity requirements (including contribution collection, contributor database development, actuarial and policy analysis, regulation of multi-pillar operations) to assess client requirements and determine h o w best they can b e met. T h i s would i n v o l v e ensur- ing that a p l a n for technical assistance i s put i nplace for r e f o r m initiatives so that client capacity i s developed. ’ “Pension Reform: Issues and Prospects for Non-Financial Defined Contribution (NDC) Schemes,” World Bank 2006, forthcoming. 136 Attachment 1 Management Response 17. Development of institutions i s a n essential p a r t of most pension reforms and assess- ing capacity-building requirements are k e y t o the development of a successful reform. Management now addresses overall issues of capacity building in the context of results- based CASs, according t o the priorities set out i ncountry-owned plans such as PRSs. Ca- pacity, along with governance and country results frameworks, are standard subjects cov- ered during Management review of draft CASs. Capacity-building plans are prepared by clients with support from the Bank and other development partners. To the extent that the Bank i s financing technical assistance for s u c h capacity building, Management will continue to require that plans are put in place consistent with project implementation needs. Man- agement would not see a checklist as necessarily effective, given the need to adapt B a n k support to country conditions. Moreover, the Bank can only review capacity-building tech- nical assistance programs to the degree that it i s requested to do so by country authorities. Recommendation 3. Conduct Research on Outstanding Issues: Ensure that ade- quate analysis i s conducted o n k e y issues such as income of the aged, the i m p a c t of corruption and governance o n the feasibility o f effective pension regulation, meth- ods to foster competition among pension funds, guidelines for investment alloca- tion, the design of non-contributory systems, a n d ways in w h i c h capital markets de- velop, as w e l l as supportive research that can provide cross-country evidence on these topics. 18. Management agrees that adequate analytical work i s essential to i t s work supporting the strengthening and r e f o r m of pension systems. However, there are many other compet- ing priorities for country analytic work. If CASs identify pension r e f o r m as a central element of the Bank's support efforts, Management will ask C A S teams to review the knowledge base, including analytic work done by the client country a n d other development partners, a n d to address the issue o f covering identified gaps. Management considers this the now- current practice a n d therefore plans n o further action on this recommendation. With r e g a r d to cross-country analytic a n d research topics, a task force representing H D N , FSE, PREM, a n d DEC n e t w o r k and Regional staff will r e v i e w needs and set priorities for consideration in the work p r o g r a m process of the relevant units. K e y issues such as development of im- p r o v e d pension regulation a n d oversight appropriate to country conditions; alternative de- sign options for unfunded contributory pension schemes; policy options to ensure mini- mum income support a n d poverty risk reduction for the elderly including non-contributory schemes; a n d 2nd pillar fund performance and indicators will b e priorities for consideration for further research. Recommendation 4. I m p r o v e Internal and External Coordination: (a) Develop a process t o ensure that cross-sector issues are considered including financial issues identified by the FSAP and maintain closer coordination between the Development Economics vice presidency, the Networks, sector units, and country units; a n d (b) Develop a strategy to p l a y a greater role in consensus building among stakeholders, inparticular, international organizations and client agencies. 19. InternaZ Coordination. Management agrees with the importance of internal and ex- ternal coordination. Ina n effort at strengthening internal coordination, the task force out- l i n e d above representing HDN, FSE, PREM a n d DEC n e t w o r k a n d Regional staff will meet n the development of central and coun- periodically t o review inter-sectoral collaboration i 137 Attachment 1 Management Response try-level outputs and strategies and i n the context of the analytical and research priorities n o t e d above. The recently initiated FSAP Updates a n d ESW focused on pensions, insurance a n d capital markets wiU play a critical role in t h i s process. 20. External Engagements. The Bank anchors i t s Country Assistance Strategies in a country’s own v i s i o n for i t s development as defined in a Poverty Reduction Strategy or other country-owned process. With t h i s as the framework for country level engagements, staff will continue t o actively w o r k with country authorities a n d coordinate with other de- velopment partners. The Bank will stay engaged with other international organizations a n d bilateral donors a n d creditors i n discussing alternative approaches to pension reform. Management fully agrees to measures to i m p r o v e coordination and t o address differences a n d commonalities inreform proposals. 21. Attachment. Attached to this Management Response i s a n annex containing detailed responses in the Management Action Record matrix 138 Attachment I Management Response Major monitorable OED recommendations Management response requiring a response Develor, Guidelines to Design Pension Reforms and Pay f i e Bank will continue to evaluate pension systems against Greater Attention to Parametric Reforms country objectives and recommend reforms (including pa- a. Pay greater attention to parametn’c reforms to rametric and/ or structural reforms) appropriate to country ensure fiscal sustainability and to the macroeco- conditions. Management fully agrees with the importance nomic, financial, a n d institutional sector precondi- of guiding criteria for establishing privately-managed sec- tions necessary fw a multi-pillar reform. This w o u l d ond pillar reforms a n d will establish a working group repre- involve preparing a n d implementing guidezines sented by HDN, FSE, PREM a n d D E C to develop a diagnos- to ensure well-tailored assistance to country tic framework. D r a f t guidelines will be produced within one conditions and consistent policy prescriptions year, completing Management‘s commitment with regard t o including statistical indicators and in-depth as- this recommendation. sessments. I b. Be mme realistic in presenting the benefits of the Management agrees with the importance o f realistically pre- secondary objectives o f pension reform in dialogue senting the benefits of the secondary objectives of savings with client countries, as there i s insufficient em- growth and capital market development. Measures t o sup- pirical evidence to support the claims that port these objectives must be fully consistent with a hierar- funded systems have or can improve savings and chy of measures supporting a strategy towards poverty alle- capital market development. viation. This issue will be covered under the above guidelines. I Build C/ien Caaacitv t Management n o w addresses overall issues of capacity build- ing i nthe context o f results-based CASs, according t o the c. Develop a checklistfor client capacity require- priorities set out in country-owned plans such as PRSs. Ca- ments (including contribution collection, con- pacity, along with governance and country results frame- tributor database development, actuarial and works, are standard subjects covered during Management policy analysis, regulation of multi-pillar opera- review of draft CASs. Capacity-building plans are prepared tions) to assess client requirements and deter- by clients with support f r o m the Bank and other develop- mine h o w best they can be met. T h i s would in- ment partners. To the extent that the Bank is financing tech- volve ensuring that a plan for technical assistance i s nical assistance for such capacity building, Management will put in place for reform initiatives so that client ca- continue to require that plans are put in place consistent pacity is developed. with project implementation needs. Management w o u l d n o t see a checklist as necessarily effective, given the need t o adapt Bank support to country conditions. 139 Attachment 1 Management Response Major monitorable OED recommendations Management response -equiring a response :onduct Research on Oufstandinu Issues vlanagement agrees that good analytic w o r k i s important i n ;upporting the strengthening and reform o f pension sys- d. Ensure that adequate analysis is conducted on .ems. However, there are many competing demands f o r key issues such as income of the aged, the impact i m i t e d resources for country analytic work. I f CASs iden- o f corruption and governance o n the feasibility lfy pension reform as a central element of the Bank's sup- o f effective pension regulation, methods to fos- 2ort efforts, Management will ask C A S teams to r e v i e w the ter competition among pension funds, guide- mowledge base, including analytic w o r k done by the client lines for investment allocation, the design of :ountry and other development partners, and t o address the non-contributory systems, and ways i n which ssue of covering identified gaps. Management considers capital markets develop, as w e l l as research of- h i s the now-current practice and plans n o further action o n fering cross-country evidence o n these topics. h i s recommendation. With regard t o cross-country analytic md research topics, the same task force established t o strengthen coordination will review needs a n d set priorities for consideration i n the w o r k program process of the rele- vant units. With that discussion, Management will consider this action complete. Improve lnternal and External Coordination As noted above, i na n effort at strengthening internal coor- dination, a task force representing HDN, FSE, PREM, a n d e. Develop a process to ensure that cross-sector is- DEC network a n d regional staff will meet periodically t o re- sues are considered including financial issues view inter-sectoral collaboration, notably i nthe context of identified by the FSAP a n d maintain closer co- the analytical a n d research priorities noted above. ordination between the Development Econom- ics vice presidency, the Networks, sector units, and country units. Staff will continue t o actively w o r k with country authorities f. Develop a strategy to play a greater role i n con- and coordinate with other development partners i nthe con- sensus building among stakeholder, in particular, text of a country's o w n vision for its development as defined other international organizations and client in a PRS or other country-owned process. The Bank will agencies. stay engaged with other international organizations i n dis- cussing alternative approaches to pension reform a n d t o ad- dress differences a n d commonalities i n reform proposals. 140 Attachment 2: Report from the Committee on Development Effectiveness (CODE) Chairman's Summary Committee on Development Effectiveness (Meeting of October 12,2005) 1. On October 12,2005 the Committee on Development Effectiveness (CODE) consid- ered the report Bank Assistance to Pension Reform and the Development of Pension Systems to- gether with the Draft Management Response. 2. Background. The World Bank early policy research report Averting the Old Age Cri- sis (1994) offered a detailed prescriptive exposition o f a multi-pillar pension framework. The Bank's strategy for pension r e f o r m formalized in Social Protection Sector Strategy: From Safify Net to Springboard (2001 - "Strutegf') builds on the three pillar approach proposed in the 1994 report built on: (1) a publicly managed, tax-financed pension scheme; (2) a privately managed, funded scheme; a n d (3) voluntary retirement savings. Strategy moves the multi- pillar proposal beyond structural prescription. The Bank took a leading role beginning i n the 1990's supporting a w i d e variety of reforms through analytical a n d advisory services a n d lending operations. During this period, the Bank assisted 68 countries with reforms of their pension systems with m o r e than 200 loans and credits. Inaddition, the Bank issued more than 350 papers and publications on pension r e f o r m and has been a leader i nsetting the stage for debate a n d knowledge gathering i nits organization of conferences a n d semi- nars, training programs, a n d the development a n d dissemination o f a computer m o d e l for projecting pension r e f o r m options 3. Findings a n d Recommendations. This report assesses the Bank's pension r e f o r m strategy and the resulting development outcomes for Bank assistance between 1985 a n d 2004, focusing o n w o r k inaugurated i nthe 1990s. The Bank's multi-pillar strategy to support pension reforms differed regionally and by country, as a result o f client concerns and Bank experience. finds that the Bank's advice has not always been effective. W h i l e formal pen- sion systems in many countries contributed to ballooning budget deficits, the Bank's preoc- cupation with fiscal sustainability tended to obscure the broader goal of pension policy- to reduce poverty a n d i m p r o v e retirement income adequacy within afiscal constraint. To im- prove this process, recommends that the Bank: (1) develop guidelines to design pension re- forms, p a y greater attention t o parametric reforms a n d de-emphasize secondary objectives of pension reform t o support savings growth a n d capital markets development; (2) build 141 Attachment 2 Chairman's Summary client capacity; (3) conduct research on outstanding issues; and (4) improve internal and ex- ternal coordination. 4. Management Response. Management found the study comprehensive in analyzing support for pension reforms although it felt that the Bank's activities during 1985-2004 w e r e increasingly less influenced by the ideas o f the 1994 report. The evolution of Bank thinking and operations i s documented i n the 2001 "Strategy" a n d the 2005 policy position paper "OZd-Age Income Suporf in the 21st Cenfurf'. Management agreed with the general thrust of the recommendations, i nparticular with those to strengthen internal coordination and the diagnostic framework for determining country readiness for privately-managed second pil- l a r reforms. InManagement's view, the report appears to overstate the role and potential impact of Bank support; a n d could have taken a broader v i e w o f the l i n k s between pension r e f o r m a n d p o v e r t y reduction. Management put a greater weight than the review on the importance of fiscal crises. Management stated that it i s important to keep i nmind that most structural reforms undertaken in Bank client countries occurred in the late 1990s w h i l e judgment over their results can only b e made after a period of 10 or m o r e years. 5. O v e r a l l Conclusions. The Committee welcomed Is evaluation of Bank activities i n pension r e f o r m w h i c h i s critical for many clients, a n d usually complex and politically sensi- tive. Speakers generally welcomed the content and quality of the report. They f o u n d the findings relevant and pertinent to the Bank's w o r k going forward, t h o u g h some speakers n o t e d that more t i m e was required to fully observe the outcomes of the reforms. The draft Management Response (MR) was also seen as thoughtful though several speakers would have preferred greater elaboration on a number of k e y aspects including the uneven distri- bution of Bank assistance a n d concentration i ntwo regions (ECA and LAC). The discussion focused m a i n l y on areas where differences between and Management were most apparent, namely: ( 1 )preconditions for supporting multi-pillar reform, especially macroeconomic sta- bility a n d fiscal sustainability; (2) relative emphasis on the so-called secondary objectives (e.g. savings mobilization or financial market development); and (3) the Bank's vision on pension within the w i d e r issue of social protection for the population a t large, i n c l u d i n g workers i nthe i n f o r m a l sector. 6. N e x t Steps. The draft Management Response will b e revised taking i n t o account the comments a n d concerns raised at the meeting, including requests for m o r e details and pre- cision in the responses. There was broad support for a w i d e r Board discussion on strategy. The following issues were raised during the meeting: 7. Preconditions for M u l t i - P i l l a r Reform. M a n y speakers stated that the Bank should not exclude countries, particularly low-income ones that do not meet the preconditions for a multi-pillar r e f o r m (macroeconomic stability and fiscal sustainability) from the scope of i t s assistance for much-needed r e f o r m programs. Members were i nfavor of the Bank engaging n a w i d e range of countries, i n c l u d i n g low-income a n d fragde economies a n d felt macro- i economic issues could b e addressed i nparallel. Management concurred with the impor- tance of a wider engagement with low-income economies a n d pointed out that the majority of Bank support has been towards first-pillar reforms a n d in a number of cases has discour- aged movements towards second-pillar reforms where conditions were not appropriate. 142 Attachment 2 Chairman’s Summary M a n y members welcomed Management’s intention to set up a working group to develop a diagnostic framework for the second pillar. A f e w members suggested the guidelines should reflect country priorities including appropriate measures w h e n preconditions necessary for a multi-pillar r e f o r m are n o t met. 8. Deemphasizing the Secondary Objectives o f Pension Reform. The general senti- m e n t was not infavor of deemphasizing the secondary objectives (promotingsavings growth and building financial systems and capital markets) as suggested by . clarified i t s v i e w that these secondary objectives were indeed important but they should not b e over- emphasized as they might have been i n the past. 9. Focus of W o r l d B a n k Activities. There was b r o a d recognition o f the complexity of pension reforms and their impact on macroeconomic and fiscal stability as w e l l as long-term sustainable economic growth. Some members noted that the Bank played a valuable r o l e i n pension reforms w h i l e others felt that technical assistance has been insufficient and discon- tinuous. Members generally agreed with Management that the Bank’s involvement i n pen- sion reforms was triggered by serious fiscal crises w h i c h unsustainable pension schemes contributed to i nm a n y countries. Some speakers indicated that the Bank should r e m a i n en- gaged i n this area because many developing countries were undertaking pension reforms and faced the challenge of m a k i n g the systems financially sustainable. The need for im- p r o v e d internal and external coordination, communication strategies, a n d monitoring and evaluation of r e f o r m progress was also noted. Speakers emphasized that political economy including governance issues should b e given due attention. Members were disappointed that Bank assistance to reforms did n o t necessarily l e a d to expanded pension coverage for the working population; some Members encouraged the Bank to b e m o r e proactive i nensur- ing the provision of safety nets for people i n the i n f o r m a l sector. noted two zuays in which public pension programs can provide safety nets for old-age income security for the informal sector through: (2) a tax-financed, non-contributo y pension; and (2) a means-tested pension for the aged and the disabled. 10. Capacity Building a n d Research. Some members n o t e d that assistance to countries should b e tailored to their needs, particularly capacity building, and encouraged the Bank to evaluate potential simplification measures for users of i t s pension r e f o r m options simulation model. The Bank’s analytical w o r k should cover fiscal sustainability, alternatives to r e f o r m pay-as-you-go (PAYG) systems, and the informal sector, among other relevant issues. Man- agement pointed out the tradeofis in research within its budget envelope. W h i l e noting that the majority of projected outcome ratings have been satisfactory, members observed that re- search h a d been occasionally incomplete, and h a d not always translated i n t o effective opera- tions. M a n y speakers proposed m o r e research to review the impact of HIVIAIDS on pen- sion systems, as w e l l as impact that pension r e f o r m h a d h a d on elderly population groups. 11. Country a n d Regional Perspective. While noting the uneven distribution of B a n k assistance -mainly concentrated i n ECA and LAC, a f e w members n o t e d that the Bank should address the needs of many countries i n underserved regions. A question w a s raised on how to expand the safety n e t for a large section of the population w h i c h does not have any pension benefit, inparticular i nlow-income African countries with high levels of pov- erty. A f e w members suggested that addressing pension reforms i n CASs will assure sys- 143 Attachment 2 Chairman’s Summary tematic Bank support, including adequate Analytical and Advisory Activities (AAA). Man- agement suggested the importance of comparing the relative vulnerability a n d social r i s k management options afforded to different groups i n order to determine whether scarce re- sources should support the elderly poor or perhaps be better allocated to other m o r e vul- nerable populations (such as children and disabled). 12. Other Suggestions. Some members felt that review could have covered: (1)a broader v i e w of the links between pension r e f o r m a n d poverty reduction rather than look- ing only a t the coverage a n d level of pensions p a i d to elderly; (2) differences between IMF a n d World Bank in their views on income a n d social security revenue collection; a n d (3) la- b o r market issues. Regarding the revision of Management Response some members sug- gested the following changes: (1)more clarity, especially inareas where there i s disagree- m e n t with findings a n d recommendations; (2) more detailed analysis of recommendations on building client capacity a n d conducting research on outstanding issues; and (3) m o r e in- formation on how lessons of experience have l e d t o changes i n the Bank‘s approach thereby leading to changes i n Bank lending operations. A question was also raised about a n y or Management analysis o f the supply side of the sector. Chander MohanVasudev, Chairman 144 References Aaron, H.J. (1966). ”The Social Insurance Barr, N. (ed.) (1994). “Labor Markets and Social Paradox,” Canadian Journal of Policy i nCentral and Eastern Europe: Economics and Political Science, Vol. 32, The Transition and Beyond.“ World pp. 371-374. Bank in association with the London School of Economics and Political Aaron, Henry J, Joseph A. Pechman, and Science, Oxford University Press, N e w Michael K. Taussig (1968). “Social York, NY. Security: Perspectives for Reform.” Washington, D.C.: The Brookings Barrand, P. S. Ross and G. Harrison. (2004). Institution. ”Integrating a Unified revenue Administration for Tax and Social Alier, M. and D. Vittas (1999). “Personal Contribution Collections: Experiences Pension Plans and Stock Market of Central and Eastern European Volatility,” Development Research Countries.” I M F Working Paper Group, World Bank (September). WP/O4/237 (December). Arnold, R.D., M. J. Gratez, and A.H. Munnell Barro, R.J. (1974). “Are Government Bonds N e t (eds.) (1998). ”Framing the Social Wealth?” Journal of Political Economy, Security Debate: Values, Politics and 82(6), 1095-1117. Economics, National Academy of Social Insurance.” Washington D.C. Bateman, Hazel (2004). “Evaluation of World Bank Assistance o n Pension Reform- Asian Development Bank (2003). “Technical South Korea.” Forthcoming Working Assistance Performance Audit Report Paper. World Bank, Washington D.C. on the Reform of Pension and Provident Funds in Selected Developing Member Beattie, R. and McGillivray, W. (1995). “A Risky Countries.” Asian Development Bank. Strategy: Reflections o n the World Bank TPA: REG2003-31. Report Averting the Old Age Crisis.” International Social Security Review, Auerbach, A., Gokhale, J., Kotlikoff. L. (1994). ~01.48,No.3-4. “Generational Accounting: A Meaningful Way to Evaluate Fiscal Bodie, Z. and 0.S. Mitchell (1996). “Pension Policy.” Journal of Economic Security in an Aging World.” In Perspectives 8/73-94. Securing Employer Provided Pensions: An International Perspective. Edited by Bailliu, Jeanine and H e l m u l t Reisen (1997). ”Do Z. Bodie, OS. Mitchell and J.Turner. Funded Pensions Contribute to Higher Pension Research Council, Univ. of Savings? A Cross-Country Analysis.“ Pennsylvania Press. OECD Development Centre Technical paper 130. Bosworth, B.P., and G. Burtless (1998). ”Population Aging and Economic Barr, Nicholas (2000). ”Reforming Pensions: Performance.” InB.P. Bosworth and G. Myths, Truths, and Policy Choices.” Burtless, eds., Aging Societies: The IMF, Working Paper WP/00/139, Global dimension. Wsington: The Washington, D.C. Brookings Institutions. 145 References Bosworth, B., Dornbusch, R. and Laban, R., eds. Paper 04-23, Brunel University, United (1994). ”The Chilean Economy: Policy Kingdom. Lessons a n d Challenges.” Washington: Davis, E. P. (2005). “The Role of Pension Funds The Brookings Institution. nEmerging as Institutional Investors i Bourguignon, Francois, Jean-Jacques Dethier , Markets”, paper presented at the Martin Cicowiez and Leonard0 Korean Development Institute Gasparini and Pierre Pestieau. (2005). conference ”Population Aging i nKorea: “What Impact Would a Minimum Economic Impacts and Policy Issues, Pension Have o n O l d Age Poverety? Seoul. Evidence f r o m Latin America.“ Francois Diamond, Peter (1977). “A Framework for Social Bourguignon and Jean-Jacques Dethier Security Analysis,” Journal of Public (The World Bank, Washington, DC.), Economics, Vol. 8, pp. 275-298. Browning, E. (1973). “Social Insurance and Diamond, P.A. (2002). ”Social Security Reform, Intergenerational Transfers,” Journal of the Lindahl Lectures. ” Oxford Law and Economics 16,215-237. University Press, Oxford, UK. Buchanan, James (1968). “Social Insurance in a Diamond, P., and Orszag, P. (2002). ”Reducing Growing Economy: A Proposal for Benefits and SubsidizingIndividual Radical Reform,” National Tax Journal, Accounts: An Analysis of the Plans 21 (4): 386-395. Proposed by the President’s Burtless, G. (2003). “What Do We Know About Commission to Strengthen Social the Risk of Individual Account Security.” Washington, D.C.: Center o n Pensions? Evidence from Industrial Budget and Policy Priorities, and N e w Countries,” American Economic York, NY: Century Foundation. Review. Diamond, P.A. and P.R. Orszag. (2004). “Saving Catalan, M., G. Impavido and A.R. Musalem Social Security: A Balanced Approach. ” (2000). ”Contractural Savings or Stock Brookings Institution Press, Washington Market Development: Which Leads?” DC. Journal of Applied Social Science Disney R, (1995).”0ccupational Pension Studies, 120 (3): 445-87. Schemes: Prospects and Reforms i nthe Chlon, A., M. Gora, and M. Rutkowski (1999). UK.” Fiscal Studies, vol. 16, No. 3, pp. “Shaping Pension Reform i n Poland: 19-39. Security through Diversity,” SP Dobronogov A., and M. Murthi (2005). Discussion Paper 9923, World Bank, “Administrative Fees and Cost of Washington D.C. Mandatory Private Pensions in Chowdhury, Abdur (2003). “Banking Reform i n Transition Economies.” Journal of Russia: Winds of Change?” BOFIT No. Pension Economics and Finance, vol. 4. 5, Institute for Economies i nTransition, (1): 31-55. Cambridge University Press. Hensinki, Finland. Douglas, Arnold, R.; Graetz, Michael; and Cichon, M. (1999). ”Notional Defined- Munnell, Alicia H., eds. (1998). Contribution Schemes: Old Wine i n “Framing the Social Security Debate: New Bottles?” International Social Values, Politics and Economics.” Security Review, Vol. 52 (4): 87-1056. Washington, DC: The Brookings Press. Davis, Philip E. and Yuwei Hu (2004). ”Is there a Feldstein, Martin (1974). “Social Security, Link between Pension-Fund Assets and Induced Retirement, and Aggregate Economic Growth?” Department of Capital Accumulation.” Journal of Economics and Finance Discussion Political Economy 82, no. 5: 905-27. 146 References Feldstein, Martin (1996), ”The Missing Piece in nCentral Fultz, E. (ed.) (2002). “Pension Reform i Policy Analysis: Social Security reform.” and Eastern Europe Volume 1 : American Economic Review, Papers and Restructuring of Public Pension Proceedings 86(1): 1-14. Schemes: Case Studies of the Czech Republic and Slovenia.” International Feldstein M. (ed.) (1998). ”Privatizing Social Labour Office, Central and Eastern Security.” The University of Chicago European Team, Budapest, Hungary. Press. Fultz, E., M. Ruck and S. Steinhilber (eds.) Ferrara, J. Peters (1982). ”Social Security: (2001). ”The Gender Dimensions of Averting the Crisis (Washington, 13G.: Social Security Reform in Central and Gat0 Institute). Eastern Europe: Case Studies of the Fields, G.S. and O.S. Mitchell (1984). Czech Republic, Hungary and Poland.” ”Retirement, Pensions, and Social Sub-regional Office for Central and Security M I T Press.” Cambridge, Eastern Europe, International Labour Massachusetts. Office, Budapest, Hungary. Fornero, Elsa and Pier Marco Ferraresi. (2004). Gill, Indermit S., Truman Packard and Juan ” Country Evaluation of Pension Yermo (2005). ” Keeping the Promise of Reforms-Albania.” Forthcoming n Latin America.” Social Security i Working Paper. World Bank, Washington, DC.: W o r l d Bank. Washington D.C. Gdlion, Colin; John Turner; Clive Bailey; and Fornero, Elsa and Pier Marco Ferraresi. (2004). Denis Latulippe (editors), 2000. “Social ” Country Evaluation of Pension Security Pensions: Development and Reforms Activities- Former Yugoslav Reform.” International Labour Office, Republic of Macedonia.” Forthcoming Geneva, Switzerland. Working Paper. World Bank, Glaessner, T. C. and S. Valdes-Prieto (1998). Washington D.C. “Pension Reform i nSmall Developing Fornero, Elsa and Pier Marco Ferraresi. (2004). Countries,” Policy Research Working ” Country Evaluation of Pension Paper No. 1983, Latin America and the Reforms Activities- Federation of Caribbean Region, W o r l d Bank/ Russia.” Forthcoming Working Paper. Washington D.C. World Bank, Washington D.C. Grootaert, C. and Braithwaite, J. 1998. ”Poverty Fornero, Elsa and Pier Marco Ferraresi. (2004). correlates and indicator-based targeting ” Country Evaluation of Pension in Eastern Europe and the former Soviet Reforms Activities- Republic of Latvia..” Union,” Policy Research Working Paper Forthcoming Working Paper. World no. 1942, World Bank, Washington, D.C. Bank, Washington D.C. Gustman A. L. and T. L. Steimeier (1986). “A Friedman, M. J. (1972). Second Lecture, i n Structural Retirement Model, Cohen, W. J., Friedman, M. J. (eds), Econometrica,” 54: 555-584. ”Social Security: Universal or Hammer, P.L., A. Kogan, and M.A. Lejeune. Selective?” Washington, DC: American (2004). “Country Risk Ratings: Enterprise Institute, 21-49. Statistical and Combinatorial Non- n Fultz, E. (ed.) (2002). ”Pension Reform i Recursive Models,” Rutcor Research Central and Eastern Europe Volume 1 : Report RRR8-2004, Rutgers Business Restructuring with Privatization: Case School, Rutgers University, Newark, NJ. Studies of Hungary and Poland.” Hanson, J., and S. Ramachandran (2005). International Labour Office, Central and “Financial Liberalization: What Went Eastern European Team, Budapest, Right, What Went Wrong?” i nWorld Hungary. Bank, Economic Growth in the 1990’s: 147 References Learning f r o m a Decade of Reform, James, E. (1992). “Income Security for Old Age: (Washington D.C.: World Bank). Conceptual Background and Major Issues.” Policy Research Working Paper, Heller, P.S. (2003). ”Who W i l l Pay? Coping Public Sector Management and Private with Aging Societies, Climate Change, Sector Development, SPS 977, World and Other Long-Term Fiscal Bank, Washington D.C. Challenges.” International Monetary Fund, Washington, D.C. Kakwani, Nanak and Kalanidhi Subbarao (2005). ”Ageing and Poverty i n Africa Holzmann, R. and R. Hinz (2005). “Old Age and the Role of Social Pensions,” Sector Income Support i n the 215t Century: An Report No. 32178 (Washington D.C.: International Perspective o n Pension World Bank). Systems and Reform.” World Bank, Washington D.C, (forthcoming). Kaufmann, Daniel, Aart Kraay and Massimo Mastruzzi, (2004). ”Governance Matters Holzmann, R., M. Orenstein, and M.Rutkowski 111: Governance Indicators for 1996- (eds.) (2003). “Pension Reform in 2002.” Washington DC.: World Bank. Europe: Process and Progress. Directions in Development”. World Kaufmann, Daniel, Aart Kraay and Massimo Bank, Washington D.C. Mastruzzi, (2005). “Governance Matters IV: Governance Indicators for 1996- Holzmann, R., I.W. Mac Arthur, and Y. Sin 2004.” Washington DC.; World Bank. (2000). ”Pension Systems i nEast Asia and the Pacific: Challenges and Kaufmann, D, A. Kraay, and P. Zoido-Lobaton Opportunities,“ SP Discussion Paper No. (1999). ”Governance Matters,” Policy 0014, Social Protection, the World Bank, Research Working Paper No. 2196, Washington D.C. Development Research Group, Macroeconomics and Growth and Holzmann, R. T. Packard, and J. Cuesta (2000). World Bank Institute, Governance, “Extending Coverage in Multi-Pillar Regulation, and Finance, World Bank, Pension Systems: Constraints and Washington D.C. Hypotheses, Preliminary Evidence and Future Research Agenda,” SP Discussion Kudat, Ayse and Nadia H.Youssef (1999). Paper No. 0002, Social Protection, World “Older People inTransition Economies: Bank, Washington D.C. An Overview of their Plight i nthe ECA Region,” ECSSD Working Paper No. 12 Holzmann, R. (1998). “A World Bank (21039), World Bank, Washington D.C. Perspective o n Pension Reform,” SP Discussion Paper No. 9807, Social Kotlikoff, Laurence J. (1988). ”Intergenerational Protection, World Bank, Washington Transfers and Savings.” i n”What D.C. Determines Savings?” Cambridge, MA: MIT Press, pp. 141-166. Impavido, Gregoria; Albert0 R. Musalem and DimitiVittas (2002). “Contractual Kotlikoff, L.J (1994). ”Rethinking the World nCountries with a Small Savings i Bank’s Social Insurance Analysis.” ESP Financial Sector,” Policy Research Discussion Paper Series, no. 31. Working Paper No. 2841 (Washington Leimer, D.R., and S.D. Lesnoy (1982). ”Social D.C.: World Bank). Security and Private Saving: N e w Time International Federation of Pension Fund Series Evidence,” Journal of Political Administrators (2003). “Pension Economy. Reforms: Results and Challenges,” Lindeman, D., M. Rutkowski, and 0. Corporaci6n de Investigacibn, Estudio y Sluchynskyy (2000). “The Evolution of Desarrollo de l a Seguridad Social, nEastern Europe and Central Systems i Santiago, Chile. Asia: Opportunities, Constraints, 148 Refe rences Dilemmas and Emerging Practices.” Discussion Paper Series No. 0009 World Bank, Washington, D.C. (Washington D.C.: W o r l d Bank). Mackenzie, G.A. (2002). ”The Role of Private Palmer, Edward (2004). ” Country Evaluation of Sector Annuities Markets i nan Pension Reform Activities-Bulgaria.” Individual Accounts Reform of a Public Forthcoming Working Paper. World Pension Plan.” I M F Working Paper, Bank, Washington D.C. International Monetary Fund, Palmer, Edward (2004). “ Country Evaluation of WP/02/161. Washington D.C. Pension Reform Activities-Hungary .” Martino, J.S. (2003). “Evaluation of World Bank Forthcoming Working Paper. Worid Assistance to Pension Reform i n Bank, Washington - D.C. Argentina.” Forthcoming Working Palmer, Edward (2004). ” Country Evaluation of Paper. World Bank, Washington D.C. Pension Reform Activities-Kazakhstan.” Martino, J.S. (2003). ”Evaluation of World Bank Forthcoming Working Paper. World Assistance to Pension Reform i n - D.C. Bank, Washington Uruguay.” Forthcoming Working Palmer, Edward (2004). ” Country Evaluation of Paper. World Bank, Washington D.C Pension Reform Activities-The Kyrgyz Muller, K. (2003). ”Privatising Old-Age Security: Republic.” Forthcoming Working Latin America and Eastern Europe Paper. World Bank, Washington D.C Compared‘” Edward Palmer, Edward (2004). ” Country Evaluation of Northampton MA. Pension Reform Activities Regional Musalem, A. R. and R. J. Palacios (eds.) (2004). Summary: Albania, Bulgaria, Croatia, “Public Pension Fund Management: Hungary, Kazakhstan, The Kyrgyz Governance, Accountability, and Republic, Latvia, FYR Macedonia and Investment Policies.” World Bank, Russia.” Forthcoming Working Paper. Washington D.C. World Bank, Washington D.C. Organisation for Economic Co-operation and Pechman, J.A., H.Aaron and M.K.Tausig (1968). Development. (2003). “Reforming Public “Social Security: Perspectives for Pensions: Sharing the Experiences of Reform.” The Brookings Institution, Transition and OECD Countries.” Washington, D.C. OECD, Paris France. Piggott, John (2004). ” Country Evalution of . (2001). ”Towards Asia’s Pension Reform Activities Regional Sustainable Development: The Role of Summary: Asia.” Forthcoming Working Social Protection.” OECD, Paris, France. Paper. World Bank, Washington D.C. Orszag, P.R. and J.E. Stigliz (2001). “ Rethinking Piggott, John and Lu Bei (2004). ”World Bank- Pension Reform: Ten Myths about Social Country Evaluation of Pension Security System,” i n R. Holzmann and Activities: People’s Republic of China.” J.E. Stiglitz, eds., “New Ideas about Old Forthcoming Working Paper. World Age Security.” World Bank, Washington Bank, Washington D.C. D.C. Queisser, M., C. Bailey, and J. Woodall (1997). Quadangno, J. (1988). ”The Transformation of ”Reforming Pensions i n Zambia: An Old Age Security: Class and Policies in Analysis of Existing Schemes and the American Welfare State.” The Options for Reform, Vol. I . Policy University of Chicago Press: Chicago, Research Working Paper, No. WPS 1716 IL. (Washington DC: W o r l d Bank). Palacios, Robert and Montserrat Pallares- Ramachandran, S., and I .Kessides (2005). Miralles (2000). “International Patterns “Privatization and Deregulation: A Push of Pension Provision,” Social Protection Too Far?” i n W o r l d Bank, Economic 149 References G r o w t h in the 1990’s: Learning f r o m a Reform in Brazil.” Forthcoming Decade of Reform (Washington D.C.: Working Paper. W o r l d Bank, W o r l d Bank). Washington D.C. Reynaud, E, (ed.) (2000).”Social Dialogue a n d Va1di.s-Prieto, Salvador (2004). ”Evaluation of Pension Reform.” International Labour W o r l d Bank Assistance t o Pension Office: Geneva, Switzerland. Reform in Mexico.” Forthcoming Working Paper. W o r l d Bank, Riesen H. and J. Bailliu (1997). ’Do Funded Washington D.C Pensions Contribute to Higher Aggregate Saving? A Cross-country Valdes-Prieto, Salvador (2005). ”Evaluation of Analysis.” OECD Development Centre Pension Related Activities: Regional Technical Papers, N o . 130, OECD, Paris. Summary of L a t i n America 1992-2002.” Forthcoming W o r k i n g Paper. W o r l d Rofman, Rafael (2002). ”Evaluation of W o r l d Bank, Washington D.C. Bank Assistance to Pension Reform-The Case of Peru.” Forthcoming Working Vittas, Dimitri (1998). “Regulatory Controversies Paper. W o r l d Bank, Washington D.C. of Private Pension Funds,” W o r l d Bank Policy Research W o r k i n g Paper WPS Samuelson, P. (1958). ”An Exact Consumption- 1893 (Washington D.C.: W o r l d Bank). loan M o d e l of Interest.” Journal of Political Economy, vol. 66. Vittas, D. (1998). ”Institutional Investors and Securities Markets: W h i c h Comes Samwick, Andrew A. (1999). “Social Security First?” Policy Research Working Paper No. Reform in the U n i t e d States.” National 2032, Development Research Group, Tax Journal, 52 (December), 819 - 842. Finance. Washington D.C.: W o r l d Bank Schieber, Sylvester J., and John B. Shoven, Walker, E. and F. Lefort (2002). Pension Reform (1999). ”The Real Deal.” N e w Haven: a n d Capital Market: A r e there any Yale University Press. (Hard) Links?” Social Protection Steurle, C.E. and J.M. Bakija (1994). “Retooling Discussion Paper N o . 0201 (Washington Social Security for the 215‘ Century: D.C.: W o r l d Bank). Right and W r o n g Approaches to Weaver, Carolyn L. (1981). ”Understanding the Reform.’’ Urban Institute Press, Sources and dimensions of Crisis in Washington D.C. Social Security: A First Step T o w a r d Thompson, Lawrence. (1998). “Older and Wiser; Meaningful Reform.” Fiscal Policy The Economics of Public Pensions.” Council, Washington D.C. Washington, DC: U r b a n Institute Press. Whitehouse, E d w a r d (2000). “ H o w Poor A r e Thompson, L.H. (1983). “The Social Security the Poor? A Survey o f Evidence of 44 Reform Debate”, Journal o f Economic Countries,” Social Protection Discussion Literature, Vol. 21:4: 1425-1467. Paper No. 0017 (Washington D.C.: W o r l d Bank). Turner, J.A., and N. Watanabe (1995). ”Private Pension Policies in Industrial Countries: W o r l d Bank (2005). “Old Age Income Support in A Comparative Analysis.” Michigan the 21st Century: The W o r l d Bank’s Upjohn, Kalamazoo. Perspective o n Pension Systems and Reform.” Washington D.C.: W o r l d Valdes-Prieto, Salvador (2004). ”Evaluation o f Bank. W o r l d Bank Assistance t o Pension Reform in Bolivia.” Forthcoming . (2004). ”China: Country Assistance Working Paper. W o r l d Bank, Evaluation.” Report No. 29734. Washington D.C. Independent Evaluation Group. Washington D.C.: W o r l d Bank. Valdes-Prieto, Salvador (2004). ”Evaluation of W o r l d Bank Assistance t o Pension 150 References . (2004). “Approach Paper: Review . (2001). “New Ideas about Old Age of Bank Assistance to Pension Reform Security.” Washington D.C.: W o r l d and the Development of Pension Bank. System.” Independent Evaluation . (1994). “Averting the Old Age Group. Washington D.C.: World Bank. Crisis. Oxford: Oxford University. . (2001). ”Social Protection Sector World Economic Forum (2000). ”The Africa Strategy: From Safety Net to Competitiveness Report 200012001.” Springboard.” Washington D.C.: World Oxford University Press, N e w York. Bank. 151