MONGOLIA POLICY OPTIONS FOR PENSION REFORM Report No. 68526-MN January 20, 2011 Human Development Unit East Asia and Pacific Region 1 List of Acronyms ADB Asian Development Bank CPI Consumer Price Index DB Defined Benefit DC Defined Contribution DSA Debt Sustainability Analysis FDC Funded Defined Contribution GAPP Generally Accepted Principles and Practices GDP Gross Domestic Product IMF International Monetary Fund MDC Matching Defined Contribution MLS Minimum Living Standard MNT Mongolian Tugrits MSWL Ministry of Social Welfare and Labor NDC Notional Defined Contribution PIF Pension Insurance Fund PIS Pension Insurance Scheme PROST Pension Reform Options Simulation Toolkit SIGO Social Insurance General Office SWF Social Welfare Fund UIF Unemployment Insurance Fund WIF Work Injury Fund Regional Vice President Pamela Cox (EAPVP) Country Director: Klaus Rohland (EACCF) Country Manager: Coralie Gevers (EACMF) Sector Director: Xiaoqing Yu (EASHD) Acting Sector Manager: Hana Palackova Brixi (EASHS) Task Team Leader: Mark Dorfman (HDNSP) 2 Policy Options for Pension Reform Acknowledgements This report was managed and written by Mark Dorfman (Senior Economist, HDNSP) based on the inputs from Tungalag Chulun (Operations Officer, EASHE), Tatyana Bogomolova (Social Protection Specialist, HDNSP), Lawrence Thompson (Consultant), Nasir Whaind (Consultant), Oleksiy Ivaschenko (Senior Economist, EASHS) and Michele Zini (Consultant, EASHS). The report benefited from the inputs of peer reviewers Robert J. Palacios (Pensions Team Leader, HDNSP) and Lawrence Thompson (Consultant). The report was prepared under the guidance of Philip O’Keefe (Country Sector Coordinator, Human Development, EASHS), Xiaoqing Yu (former Sector Manager, EASHD) and Emmanuel Jimenez (former Sector Director, EASHD). It is based on visits to Ulan Bataar in October and December, 2010, May and September, 2011. The study is one of several outputs of the Mongolia Social Protection Programmatic Technical Assistance program (Oleksiy Ivaschenko, Task Manager). The study was assisted by resources of the Korean Trust Fund for Poverty Reduction and Socio-economic Development. The team would like to thank the Ministry of Labor and Social Welfare, the Social Security Insurance Government Office, the Ministry of Finance, the Mongolian Confederation of Trade Unions and the Mongolian Confederation of Employers for providing important inputs to the study. The team would also like to thank those Members of Parliament and their staffs who provided invaluable inputs to the Bank team. 3 Policy Options for Pension Reform Table of Contents EXECUTIVE SUMMARY ..................................................................................................................... VIII I. BACKGROUND AND OBJECTIVES.....................................................................................................1 II. DESCRIPTION AND EVALUATION OF PENSION SYSTEM DESIGN .........................................................3 A. Description of Elderly Social Assistance and Contributory Pension Insurance 3 B. Demographic Context - Aging 5 C. The Economic Context 8 D. A Growing Fiscal Burden 9 E. Pension Insurance Scheme Weaknesses and Reform Needs 11 III. OBJECTIVES AND OPTIONS FOR A DESIGN ARCHITECTURE ..............................................................19 A. Objectives and Design Considerations 19 B. Tier I – Targeted Social Pensions 22 B1. Rationale for a Targeted Social Pension .................................................................. 22 B2. Recommendation, simulated benefits and fiscal costs ............................................. 23 C. Tier II – Contributory Pension Insurance 29 C1. Pension Insurance for Old Age Retirement ............................................................. 29 C1.1. Modifying Key Parameters in the existing NDC Scheme ................................ 29 C1.2. Evaluation of a defined-benefit scheme ........................................................... 32 C1.3. Evaluation of a Points scheme.......................................................................... 34 C1.4. Evaluation of a Funded Defined-Contribution (FDC) scheme ........................ 34 C2. Pension Insurance for Permanent Disability ............................................................ 36 C3. Pension Insurance for Survivorship ......................................................................... 39 D. Tier III – Supplementary Voluntary Pensions Savings 40 E. Transition Arrangements 42 IV. ADDRESSING THE SPECIAL NEEDS OF HERDERS AND THE SELF-EMPLOYED......................................44 A. The Coverage Gap 44 B. Policy Options 45 B1. Mandating contributions for herders and other informal sector workers ................ 45 B2. Expanding the Voluntary Pension Program and Introducing Matching Contributions .......................................................................................................................... 47 B3. Integrating the Social Welfare Pension into a Targeted Social Pension .................. 50 B4. A recommended approach ....................................................................................... 50 V. PRE-FUNDING TRANSITIONS AND FUTURE OBLIGATIONS THROUGH A PENSION RESERVE FUND ........55 A. Conceptual Considerations 55 B. Pre-Funding Transitions and Future Obligations through a Pension Reserve Fund55 C. Recommendations 59 VI. GOVERNANCE AND INSTITUTIONAL REFORM ................................................................................60 VII. SUMMARY RECOMMENDATIONS AND POSSIBLE REFORM PROCESS..................................................62 A. Objectives 62 B. Overall Design 62 C. Tier I – Targeted Social Pensions 62 D. Tier II – Contributory Pension Insurance scheme 62 E. Tier III – Voluntary Occupational and Individual Pensions Savings Arrangements 63 4 Policy Options for Pension Reform F. Addressing the Special Needs of Herders and the Informal Sector 63 G. Pre-Financing Future Pension Liabilities 63 H. A Reform Process 63 VIII. CONCLUSIONS............................................................................................................................66 BIBLIOGRAPHY ..............................................................................................................................68 APPENDIX 1: SOCIAL INSURANCE PENSION PARAMETERS ...............................................................71 APPENDIX 2: GLOSSARY ...............................................................................................................74 APPENDIX 3: OPTIONS FOR PRE-FINANCING PENSION LIABILITIES ..................................................78 APPENDIX 4: SUMMARY OF GENERALLY ACCEPTED PRINCIPLES AND PRACTICES (GAPP) – THE “SANTIAGO PRINCIPLES� ..........................................................................................83 APPENDIX 5: SIMULATION OF REFORM OPTIONS ............................................................................85 A. Parametric Reforms of the NDC Scheme 85 B.. Parametric Reforms of the NDC Scheme 87 C. Extension of a Defined-Benefit Scheme to Post-1960 Cohorts 92 D. Establishment of a Funded Defined Contribution Scheme for Post-1960 Cohorts 95 APPENDIX 6: ACTUARIAL PROJECTION METHODOLOGY ...................................................................98 A. Introduction 98 B. Demographic Projections and Assumptions 99 B1. Country Population .................................................................................................. 99 C. Economic Assumptions 101 C.1 Projected GDP Growth .......................................................................................... 101 C.2 Productivity Growth............................................................................................... 102 C.3 Projected Inflation .................................................................................................. 102 D. Labor Force Projections 102 D.1 Labor Participation Rates ....................................................................................... 102 D2. Unemployment Rates ............................................................................................. 103 E. Pension System Contributors 103 F. Pension System Beneficiaries 104 F1. Old-Age Pensioners by Gender/Age Distribution ................................................. 104 F2. Disabled Pensioners by Gender/Age Distribution ................................................. 105 F3. Survivor and Orphan Pensioners by Gender/Age Distribution.............................. 105 G. Pension System Cash-Flows 105 G.1 Pension System Cash In-Flows .............................................................................. 106 G.2 Pension System Cash Out-Flows ........................................................................... 107 5 Policy Options for Pension Reform Tables Table 1: Summary of Reform Needs and Proposed Policy Options .................................... xiv Table 2: Recipients and Benefits for Social Welfare Pensions for Elderly, Disabled and Survivors ...................................................................................................................... 3 Table 3: Social Insurance Benefit Expenditures .................................................................... 10 Table 4: Average Wage, Minimum Wage and Minimum Living Standard ........................... 12 Table 5: Social Insurance Contribution Rates ....................................................................... 15 Table 6: Stylized Illustration of Factor Reduction of Targeted Social Pension................... 25 Table 7: Possible Benefit Enhancement to Smooth the Transition for Pre- to Post-1960 Cohorts ....................................................................................................................... 43 Table 8: Social Insurance Contributors, Beneficiaries, Economically Active Population .. 44 Table 9: Summary of Impact of Baseline and Proposed Reform Scenarios ........................ 86 Table 10: Projected Life Expectancy at Various Ages by Gender ......................................... 100 Table 11: Assumed Real GDP, Inflation and Wage Growth .................................................. 102 Table 12: Assumed Labor Participation Rates ...................................................................... 103 Table 13: Average Years of Contribution for New Old-Age Retirees ................................... 108 Table 14: Average Years of Contribution for New Disability Retirees ................................ 109 Table 15: Defined Benefit and NDC Scheme Minimum Pensions......................................... 110 Figures Figure 1. Total Fertility and Life Expectancy at Birth 1950-2005 ....................................... 5 Figure 2. Pension Contributors, Beneficiaries and Economically Active Population ......... 6 Figure 3. Age Structure of the Population – 2010-2050 ....................................................... 7 Figure 4. Projected Life Expectancy and Normal Retirement Age ....................................... 7 Figure 5. Population Pyramids – 2008-2058 ......................................................................... 8 Figure 6. Social Insurance Contributions and Benefits ....................................................... 10 Figure 7. Projected Baseline Defined Benefit and NDC Current Balances......................... 11 Figure 8. Baseline Defined Benefit and NDC Replacement Rates ...................................... 13 Figure 9. Retirement Age for Pension Insurance Schemes in East Asia, Europe and Central Asia ............................................................................................................. 14 Figure 10. Comparative Contribution Rates for Pension Systems in East Asia, Europe and Central Asia ............................................................................................................. 16 Figure 11. Tradeoffs in Achieving Core Objectives of Old Age Income Protection ............. 20 Figure 12. A Three Tier Approach to Old Age Income Protection ........................................ 21 Figure 13. Summary of Design Considerations for Tier I Targeted Social Pensions .......... 24 Figure 14. Projected Fiscal Costs of a Targeted Social Pension ............................................ 26 Figure 15. Projected Benefits of a Targeted Social Pension.................................................. 27 Figure 16. Stylized Illustration of Combined Benefits of Tier I and Tier II Schemes .......... 28 Figure 17. Stylized Example of Replacement Rates from Combined Tier I and Tier II Schemes................................................................................................................... 28 Figure 18. Reform Scenario (2E): Removal of the Minimum Pension, Increase in Retirement Age, Increase in Contribution Rates and Adoption of Actuarially Fair Annuity Factor ................................................................................................ 31 6 Policy Options for Pension Reform Figure 19. Reform Scenario (2E): Removal of Minimum Pension, Increase in Contribution Rates and Adoption of Actuarially Fair Annuity Factor ...................................... 32 Figure 20. Reform Scenarios 3A,B,C: Establishment of Defined-Benefit and Points Schemes for Post-1960 Cohorts ............................................................................ 33 Figure 21. Reform Scenarios 3A,B,C: Establishment of Defined-Benefit and Points Schemes for Post-1960 Cohorts ............................................................................ 33 Figure 22. Proposed Occupational and Personal Pension Arrangements ........................... 42 Figure 23. Possible Parameters of a Matching Defined-Contribution Scheme ................... 52 Figure 24. Stylistic Examples of Monthly Benefits produced by Matching Contributions . 53 Figure 25. Potential Elements and Sequencing of a Policy Reform Process ....................... 65 Figure 26. Scenario 2A: Removal of the Minimum Pension Guarantee for Post-1960 Cohorts (annual current balance as a % of GDP) ................................................ 87 Figure 27. Scenario 2A: Removal of the Minimum Pension Guarantee for Post-1960 Cohorts (replacement rates) ................................................................................. 88 Figure 28. Scenario 2B: Removal of the Minimum Pension and Increase in Ret. Age ........ 89 Figure 29. Scenario 2B: Removal of the Minimum Pension and Increase in Ret. Age ........ 89 Figure 30. Scenario 2C: Removal of the Minimum Pension and Increase in Contribution Rate .......................................................................................................................... 90 Figure 31. Scenario 2C: Removal of the Minimum Pension and Increase in Contribution Rate .......................................................................................................................... 90 Figure 32. Scenario 2D: Removal of the Minimum Pension Guarantee, Increase in Ret. Age and Increase in Contribution Rates ...................................................................... 91 Figure 33. Scenario 2D: Removal of Minimum Pension, Increase in Ret. Age and Increase in Contribution Rates ............................................................................................. 91 Figure 34. Scenario 2E: Removal of the Minimum Pension Guarantee, Increase in Ret. Age, Increase in Contribution Rates and Actuarially Fair Annuity Factors ............... 92 Figure 35. Scenario 2E: Removal of Minimum Pension, Increase in Retirement Age, Increase in Contribution Rates and Actuarially Fair Annuity Factors ............... 92 Figure 36. Scenario 3A: Reinstating Pre-1999 PAYG Defined Benefit Scheme ................... 93 Figure 37. Scenario 3A: Reinstating Pre-1999 PAYG Defined Benefit Scheme ................... 93 Figure 38. Scenarios 3B,C: Conversion to DB Scheme with Parametric Changes ............... 94 Figure 39. Scenarios 3B,C: Conversion to DB Scheme with Parametric Changes ............... 94 Figure 40. Scenarios 4A,B,C: Converting NDC Scheme to Fully Funded DC ........................ 96 Figure 41. Scenarios 4A,B,C: Converting NDC Scheme to Fully Funded DC ........................ 96 Figure 42. Scenarios 4A,B,C: Converting NDC Scheme to FDC Scheme ............................... 97 Figure 43. PROST Projection Methodology ............................................................................ 99 Figure 44. Population Projections......................................................................................... 100 Figure 45. Population Dependency Rates ............................................................................ 101 Figure 46. Projected Number of Contributors and Beneficiaries ....................................... 105 Figure 47. Projected System Dependency Ratio .................................................................. 106 Figure 48. Projected Average Replacement Rate for Existing Old Age Pensioners .......... 108 Figure 49. Projected Baseline Revenues, Expenditures and Current Balance .................. 110 7 Policy Options for Pension Reform Executive Summary This report was prepared in response to a request by the Mongolian authorities for an evaluation of the Mongolian Pension System and policy reform options. The report identifies a number of design weaknesses in the current pension insurance scheme and needs for reform in response to changes in the Mongolian economy. The report proposes design options while noting that further work is needed to refine these options and parameters based on the inputs of Mongolian policymakers. The report is organized as follows: This Executive Summary provides a brief synopsis for key policymakers of the rationale for reform and policy options considered. The main text responds to requests by the Mongolian authorities for the Bank’s assessment. It considers the design architecture and parameters of the pension insurance scheme, the design of non-contributory elderly social assistance, pension provisions for herders and the rationale and design options for a Pension Reserve Fund. The appendices provide more detailed analysis for technical staff including current pension insurance parameters; needs and policy options for herders and other informal sector workers; options for establishing a Pension Reserve Fund; and elaboration of the actuarial modeling methodology employed, key assumptions and parameters, and detailed findings. Mongolia’s economy, labor market and fiscal position are undergoing rapid changes as a result of growth in the mining sector. These changes present both challenges and opportunities for reform in pension provisions. Potential rapid growth in wages will impact pension benefits and incentives according to current parameters. Retirement benefit levels will vary substantially between different cohorts during rapid wage growth and likely result in pressure for strong indexation to limit disparities in benefits. Mining revenues present an opportunity to set aside fiscal resources to support multi-year transition arrangements needed to reform of Mongolia’s pensions insurance and elderly social assistance. Such resources can be earmarked to pre-finance transition arrangements to improve pension adequacy and sustainability such as through the constitution of a Pension Reserve Fund. Further, a portion of such resources can ensure basic income security for retirees with lowest incomes through the establishment of a targeted social pension. Key trends motivate the need for reform: • In the short-term, reforms are needed to prevent a severe reduction in benefits for new retirees in 2015 and 2020 for women and men, respectively. This abrupt reduction in benefits may understandably be seen as highly inequitable. 8 Policy Options for Pension Reform • The state subsidy for pensions and social insurance (about 2.5% of GDP) will increase substantially over time under current policies. Much of the cost increases can be attributed to projected growth of the elderly, low retirement ages and low contribution rates relative to benefits provided. • Coverage of the elderly by pension insurance will decrease and benefits for most herders and the informal sector are likely to be low because of limited participation in the pension insurance scheme and gaps in the provision of pension benefits. Substantial reductions in coverage of the labor force by pension insurance since economic restructuring in the 1990s will gradually result in a much lower proportion of the elderly with pension insurance benefits. This will particularly affect herders and the informal sector which currently have low labor force coverage levels and can contribute to income inequality of the elderly. Additional design issues contribute to weak incentives to participate and a growing fiscal burden: 1 (i) Minimum pension provisions create weak incentives for compliance, result in a majority of retirees receiving a flat minimum benefit and result in a long-run projected annual deficit for old age pensions of about 6 percent of GDP; (ii) ad-hoc indexation has created unpredictable benefits, weakened credibility and compromised basic old age income security; and (iii) disability and survivorship designs are not aligned with contributions based on actuarial principles. Key design parameters which policymakers should focus on are: • The targeted level of income replacement in retirement; • The contribution rate of employers and employees to achieve such income replacement; • The expected state subsidy by Government; and • The length of work-life and retirement to be supported by such scheme through the retirement age. This report suggests the following policy options: (i) Establish a non-contributory targeted social pension that would use Government resources to provide modest minimum subsistence income to retirees in the bottom half of the income distribution; (ii) Reform key parameters of the Pension Insurance scheme while replacing the minimum pension with a Targeted Social Pension; 1 Table 1 at the end of the Executive Summary, provides a more detailed summary of reform needs and proposed policy options. 9 Policy Options for Pension Reform (iii) Establish a stronger legal framework and oversight for private voluntary individual and occupational savings arrangements; (iv) Strengthen the incentives for herders and the informal sector to save for pensions either through mandatory or subsidized voluntary arrangements; and (v) Establish a financing plan and potentially pre-financing some of the costs of transition to put in place the above reforms. The design structure suggested by such reforms would be (Figure A): (i) Tier I – replacement of the Social Welfare Pension and the minimum pension benefits under the pension insurance scheme with a Targeted Social Pension subject to a pensions test. 2 (ii) Tier II - Modification to select parameters of the current NDC scheme (retirement age, contribution rate, indexation and annuity factor provisions) or replacement with an equivalent defined-benefit or “points� scheme aiming to achieve a similar level of income replacement redesign of disability and survivorship provisions; and enactment of matching-subsidies for voluntary savings arrangements for herders and the informal sector. (iii) Tier III - Introduction of a regulated and supervised private supplemental pensions savings framework which can accommodate the voluntary pension savings needs of individuals, herders and the informal sector. Occupational plans can also serve as a secure supplementary savings vehicle for employers seeking to provide additional pensions benefits to motivate their employees. 2 There is a tradeoff between fiscal cost and the benefit adequacy, including in the implicit indexation method adopted. A benefit reduction through a “pensions test� of income received from other retirement benefits has been proposed and, as a result, it is suggested that only retirees with below average pension insurance benefits be entitled to a targeted social pension benefit. 10 Policy Options for Pension Reform Figure A. A Three Tier Approach to Old Age Income Protection Function Objective Instrument Basic Pillar – Protection against Targeted Social Pension – Non- Tier I elderly poverty contributory targeted benefit Prevention of a material drop in Contributory Mandatory Contributory consumption due to Pillar – Tier II Pension Insurance old age, disability or death Supplementary Income supplements Supplemental occupational and Pillar –Tier III beyond pension individual savings arrangements insurance system This report evaluates four design options for post-1960 cohorts in the pension insurance scheme: (i) Retaining the current NDC scheme for but modifying key parameters including the contribution rate, retirement age, annuity factor, indexation and minimum pension; (ii) Establishing a defined-benefit scheme with similar parameters; (iii) Establishing a points scheme with similar parameters; and (iv) Replacing the current NDC scheme with a Funded Defined-Contribution scheme. 3 The report concludes that Options 1-3 would yield roughly similar results in respect of costs, risks and benefits. The parameters used in these options such as the contribution rate, accrual rate, and retirement age were found to have a much more substantial impact on pension system costs and benefits than the designs themselves. The option of a modified NDC scheme offers distinct advantages when compared to a defined- benefit scheme: (i) the scheme and supporting institutional framework is already in place; modifying parameters can be achieved within the existing legal and institutional framework; (ii) changing the structure of benefit entitlements could be viewed with concern by the contributing members; and (iii) an important feature of the NDC scheme is that the annuity factor in the benefit calculation varies both according to the age of the retiree and the year in which he or she retires. This provides an automatic adjustment to 3 A Funded Defined Contribution (FDC) scheme refers to one in which contributions and investment returns are held in an actively and competitively managed fund and the returns realized on the assets (after all fees and administrative costs) are credited to an individual’s account balance. Under the current NDC scheme, the notional interest rate is defined by law as the growth in covered wages of the past three years. Under an FDC scheme, it is possible to realize a rate of return in superior or inferior to such 3 year average wage growth and individual contributors would bear the risk of inferior returns. The additional returns in excess of wage growth would likely be modest when compared to the higher investment risks of such a scheme borne by contributors’ replacement rates. 11 Policy Options for Pension Reform changes in life expectancy which is more difficult to accommodate in a defined-benefit scheme. The fourth option to establish a Funded Defined-Contribution (FDC) scheme presents financial risks and institutional challenges which appear to outweigh the potential returns. 4 In particular, the fiscal dimensions of transitioning towards such a scheme present a substantial fiscal challenge while other enabling conditions to support such a scheme are limited including the regulation, depth and contestability of financial markets and sufficiency of governance and institutional arrangements. The report notes that Mongolia lacks an effective retirement income strategy for those who are currently uncovered by contributory pension insurance including herders, the self-employed and non-wage workers. Although the overwhelming majority of retirees receive benefits from the Pension Insurance scheme, only about half of the economically active population was covered in 2009. Herders, who are estimated to represent about a third of the labor force, and the self employed, who represent about 15 percent of the labor force, are both exempted from making mandatory contributions to the scheme. The report evaluates options for improving future retirement incomes of herders and those in the informal sector and suggests several measures including replacement of existing minimum pension and social welfare pension provisions with a Targeted Social Pension. Other measures suggested include: (a) establish regulated individual retirement savings arrangements in accordance with the development of supplementary voluntary pensions savings; (b) elaborate a plan for matching defined contributions for post-1960 cohorts participating in regulated voluntary savings arrangements; and (c) elaborate a plan for introduction of mandatory contributions to the largest herders, self-employed and informal sector workers. A plan for matching defined contributions voluntary arrangements could replace the current voluntary scheme. A recommended approach would provide a matching subsidy at least as great as the level that individuals contribute. The minimum contribution level would need to balance what is affordable to workers with the minimum contribution necessary to provide an adequate retirement benefit. Earmarking a portion of mining royalties to finance a Pension Reserve Fund can provide one instrument to manage natural resource revenues as part of a more 4 Sufficient depth, regulatory infrastructure, contestability and other enabling conditions to support such an arrangement do not exist in the domestic financial markets and may not for some time. Utilization of international markets for such reserve accumulations still presents substantial principal-agent challenges and risks for Mongolia’s legal, regulatory and supervisory structure. 12 Policy Options for Pension Reform consistent policy and governance framework. At the same time, such a Fund can earmark resources to address the short- and medium-term challenges identified by financing a well-considered policy transition process. Potential uses of such a Pension Reserve Fund could be to: (i) finance a supplementary pension benefit to smooth the reduction in income replacement for post-1960 cohorts; (ii) finance a supplementary benefit to smooth a transition for post-1960 cohorts until such time as proposed changes in retirement age and contribution rates (below) could be gradually implemented; (iii) finance a portion of the long-term impact of aging in the pension insurance scheme; and (iv) finance a targeted social pension for several years (below) as the population ages and the elderly coverage declines, particularly for herders and the informal sector. 13 Policy Options for Pension Reform Table 1: Summary of Reform Needs and Proposed Policy Options Reform Needs Proposed Design Pension Insurance scheme not well adapted Consider a three tier design which includes multiple instruments for to the diversity of characteristics and needs old age income and poverty protection. of Mongolia’s working and elderly populations. Tier I • Establish a Targeted Social Pension replacing the minimum pension • Coverage gaps for current and future guarantee for Pension Insurance & integrating the Social Welfare elderly: Pension. • Disparity in qualification between means- • Target the bottom half of the retiree income distribution and apply a tested social welfare pension and pensions-test factor reducing the social pension by a proportion of the minimum pension insurance benefit. Pension Insurance benefit received. • 1990s economic transition contributes to • Establish regulated individual retirement savings arrangements in reduction of worker coverage of herders accordance with the development of a Tier III supplementary and informal sector. voluntary pensions savings. • Elaborate a plan for matching defined contributions for post-1960 cohorts participating in regulated voluntary savings arrangements. • Elaborate a plan for introduction of mandatory Tier II contributions to the largest herders, self-employed and informal sector workers. Tier II • Abrupt reduction in benefits for those • Remove the minimum pension guarantee and replacing it with a Tier I born beginning in 1960 and other Targeted Social Pension for both pre- and post-1960 cohorts. parametric weaknesses leading to weak • Gradually increase the retirement age to 65 for men and women (for adequacy, fiscal burden and a poor link post-1960 cohorts) and eliminate early retirement for hazardous between contributions and benefits. professions so workers have a stronger incentive to work longer. • Retirement age too low to ensure a fiscal • Gradually reinstate a 19% contribution rate. sustainability and meaningful benefits for • Establish an actuarially fair annuity factor (if the NDC scheme is post-1960 cohorts. retained). • Contribution rate reduction of 2008 • The benefit formulas for disability and survivorship social insurance requires increasing state subsidies should be structured as pooled insurance products from earmarked • Disability and survivorship pensions – contributions and the benefit formula design and cost should be poor alignment between contributions and aligned with both initial and periodic actuarial valuations based on benefits, including minimum benefit changing demographics and benefit incidence. provisions. Tier III • Establish a funded-defined-contribution policy framework for A common regulatory framework and voluntary occupational and individual pensions savings arrangements. supervisory structure does not exist for • Establish a regulatory and supervisory framework for voluntary occupational and individual pensions pensions savings arrangements, including regulatory guidelines for savings arrangements investment management and the structuring of annuity arrangements. • Revise the tax treatment for licensed pensions and annuity arrangements. Mineral royalties present a potential • If a decision is taken to establish a Pension Reserve Fund, constitution opportunity to help cover a portion of future of such a fund should be part of a long-term strategy for the national pension costs and facilitate governance and effective use of mining royalties. transitions as Mongolia enacts revised • Consider the use of such a Fund to support policy transitions or future pension policies. liabilities. • Put in place proper governance arrangements for such a Fund. • Establish critical provisions for funding and withdrawal rules institutional structure; and investment policies. 14 Policy Options for Pension Reform I. Background and Objectives 1. This report identifies the needs for policy reform in the current provision of old age income and poverty protection and utilizes an actuarial model to simulate the costs and benefits of different reform options. The report was prepared in response to a request by the Mongolian authorities for an evaluation of the Mongolian Pension system, and projections of the impact of reform options on system and fiscal finances and benefit adequacy. The report identifies a number of weaknesses in the current policy design of the pension insurance system and social welfare pension scheme that need to be addressed to strengthen old-age income security for Mongolia’s population. The report proposes design options noting that additional thinking is needed by the authorities to refine the policy choices and integrate these choices into a complete reform program. 2. This report responds to a request by the authorities for the World Bank to undertake the following assessment: 5 (i) Evaluate options and design parameters for a non-contributory social pension aimed at providing pension benefits for those elderly with insufficient sources of pension income; (ii) Utilize the Pension Reform Options Simulation Toolkit (PROST) to simulate projections of the Pension Insurance scheme according to current law including the fiscal costs and income replacement effects of: (a) modifying select parameters of the current scheme; (b) replacing the Notional Defined Contribution (NDC) scheme with a Points scheme (also with modified parameters), and (c) replacing the NDC scheme with a defined-benefit scheme for pre-1960 cohorts (with modified parameters). It also responds to requests to model a reform which replaces the NDC scheme with a Funded Defined Contribution (FDC) scheme; (iii) Evaluate the needs and available instruments for pensions insurance for herders and the informal sector with a view towards improving old age income protection for these groups; and (iv) Evaluate the rationale and potential design of a Pension Reserve Fund. 3. Like many post-socialist economies, Mongolia in the early 1990s inherited an extensive social welfare system that included a State-supported mandatory scheme for old-age retirement, disability and survivorship benefits. The Mongolian pension system underwent substantial reforms in 1994, including becoming a contributory scheme. In an effort to reduce the fiscal burden and better align contributions and benefits, in 1999 an NDC scheme was established for workers born beginning in 1960. Some measures to strengthen pension administration, financing and benefits were enacted in the 2000s and 5 This report follows a report prepared in 2008 in responding to this request. See World Bank, 2008, Mongolia: Pension Reform Challenges and Reform Options. 1 in 2008 contribution rates were reduced from 19% to 14%. In 2010, some women born in 1960 began to receive early retirement benefits and the authorities became increasingly focused on the large number of women who would retire in 2015 with benefit levels substantially lower than those received by cohorts born prior to 1960. Although most elderly in 2011 were covered under the Pension Insurance Scheme, only about half of the labor force contributes to such scheme (with herders and the informal sector hardly contributing) so that in the future elderly coverage will decrease substantially. 4. Substantial structural shifts in the economy including the increasing importance of mining have presented both challenges and opportunities for reform. Rapid growth in wages will impact pension benefits and incentives according to current parameters. Benefit levels will vary substantially between retirees of different cohorts during periods of rapid wage growth. This variation will understandably lead some to seek benefit enhancements. Mining revenues also present an opportunity to pre-finance transition arrangements to develop a pension system with improved adequacy of benefits and sustainability beyond the support of State subsidies. Furthermore, such resources if allocated effectively can provide an opportunity to ensure basic income security for retirees with the lowest incomes and greatest vulnerability. 5. The report is organized as follows: The Executive Summary is designed to summarize the key challenges, policy choices and recommendations of the report for policymakers. The main text of the report addresses the requests from the authorities for a more detailed evidence-based technical analysis of the existing policies, reform needs, reform options and recommendations. The appendices provide additional detail on key issues, provide a more in-depth analysis of the actuarial projections undertaken and outline the methodology used and key assumptions. 6. The main text is organized as follows: This section provides background and introduces the report. Section II reviews the motivations for considering reforms, including changing demographics and labor market characteristics, the economic context, weaknesses in the current pension system design, weaknesses and gaps in elderly social assistance, and the substantial and growing fiscal burden posed by the pension insurance scheme into the future. Section III reviews the objectives of reform and previews design options for Tier I, Tier II and Tier III reforms. Section IV evaluates the special needs of herders and the self-employed. Section V addresses the Government’s request for an evaluation of the establishment of a Pension Reserve Fund. Section VI reviews select governance and institutional issues identified in the analysis. Section VII summarizes the report’s recommendations in bullet form and Section VIII concludes. -2- II. Description and Evaluation of Pension System Design 7. The objective of this section is to describe and evaluate the needs and rationale for reform of Mongolia’s pension provisions. It describes elderly social assistance programs and the contributory pensions insurance system, then evaluates the challenges presented by Mongolia’s aging population, rapidly changing economy and fiscal challenges of pension insurance provision. Finally, it evaluates the policy design of the pension insurance system, including disability, survivorship and tax treatment. A. Description of Elderly Social Assistance and Contributory Pension Insurance 8. Social assistance is provided for the elderly poor through the Social Welfare Pension program as well as through targeted benefits for poor households. Provisions to assist poor households through the recently enacted Social Welfare Law may likely have an important impact on poor elderly. Additional programs provide cash transfers to caregivers, other selected elderly and services to the elderly. The elderly poor also receive other forms of assistance such as nursing home services (for single elderly) and other social welfare services (e.g. meals, bathing, cleaning, and assistance in fuel collection). 9. The Social Welfare Pension program is aimed to provide assistance to men over age 60 and women over age 55 that do not qualify for social insurance pension benefits. 6 An estimated 55,345 individuals received the Social Welfare Pension benefit or benefits for disabled and survivors in 2009 (Table 2). 7 The number of these recipients has been growing each year since 2004. The Social Welfare Pension benefit at end-2010 was MNT 53,800 per month (about US$44) or about 48 percent of the Minimum Living Standard (MLS). Table 2: Recipients and Benefits for Social Welfare Pensions for Elderly, Disabled and Survivors 2004 2005 2006 2007 2008 2009 Number of Recipients 43,630 44,680 44,850 44,232 51,245 55,345 Total allocated pension (mln.tug) 7,068 7,700 11,845 14,763 22,520 25,025 Per Capita Benefit/month 13,500 14,361 22,009 27,814 36,622 37,681 Source: Ministry of Social Welfare and Labour: Facts & Figures, 2010. 10. It would be advisable to better link the Social Welfare Pension with the Pension Insurance System to improve incentives and reduce potential coverage gaps. As 6 Recipients also include single mothers/fathers with four or more children under 18, dwarf individuals who reached 16 years old, disabled over 16 years old with over 50 percent incapacitation and each child under 18 years old of family that has lost its income earner. 7 These figures represent recipients and benefits for the elderly with no caretaker or with incapable caretakers, as well as dwarfs above 16 years of age, disabled above 16 who have lost their working capacity for 50 percent or more, single parents at age 45 (women) and age 50 (men) with four or more children and children below the age of 18 who have lost the main income earner. -3- currently structured, those who are not fully or partially vested with the pension insurance system can potentially qualify for the Social Welfare Pension. While this provides an essential floor for elderly support, it also creates artificial incentives at different intervals (10 and 20 year work histories for pre-1960 cohorts and 15 years for post 1960 cohorts). Better linkages though a gradual claw-back feature discussed below can make the incentives to contribute more continuous between workers with different contribution histories and contribution levels. 11. The Pension Insurance scheme is a mandatory scheme for workers with labor contracts who contribute in order to receive old age, disability and survivorship benefits. The combined employer/employee contribution rate is 14%, having been reduced from 19% in 2008 and the retirement age when benefits can be received is age 60 for men and 55 for women, with workers in specified professions or conditions entitled to benefits as early as 10 years prior to regular retirement age. Those workers born before 1960 can qualify to receive a defined-benefit with an accrual rate of 2.25% per year for the first 20 years and 1.5% per year thereafter. The minimum pension provisions are substantial, with vested workers having 20 years of service receiving a benefit no less than 75% of the minimum wage. A lower minimum benefit is provided for workers with 10 years of contributory service. Average male workers in 2010 received a replacement rate of about 42% of the average covered wage in the first year of retirement and female workers about 33% primarily due to shorter work histories. 12. Cohorts born after January 1, 1960 receive a benefit based on a Notional Defined Contribution (NDC) benefit formula. This formula provided for a benefit based on a notional account accumulation divided by an annuity factor. The notional account accumulation is based on contributions earmarked for old age plus a notional interest rate equal to the average growth in covered wages over the past three years. The minimum pension under the NDC scheme is calculated based on a formula of 20 percent of the national average wage, plus an additional 0.5 percent of the average wage for each additional service year beyond the minimum of 15 years. As currently formulated, the minimum pension will cover the vast majority of retirees under the NDC scheme. 13. The NDC scheme was established in 1999 to reduce the fiscal burden and better align contributions and benefits. The authorities envisioned the introduction of the NDC scheme as an interim step designed towards moving to a partially-funded defined- contribution scheme in the context of largely undeveloped capital markets and limited governmental fiscal resources. The policy documents of the period set as a goal that by 2005, 3 percent of the 19% total contribution rate would be actively managed and invested in securities and, over time, more funding would be permitted. Since contribution revenues in the pension insurance fund never exceeded benefit disbursements, there was never an opportunity for such active reserve management. -4- B. Demographic Context - Aging 14. Mongolia is undergoing a very rapid aging process which will have profound implications for the labor force, economy, households and pension insurance system. Growing old age dependency rates result from substantial declines in fertility rates and gradual increases in life expectancy at retirement age (Figure 1). Fertility rates fell from 7.5 children per woman in the 1970s to 2.1 children per woman in the 2000s while life expectancy at birth increased from 43.5 years in 1950 to 67.3 years in 2010. Figure 1. Total Fertility and Life Expectancy at Birth 1950-2005 8 80 Total Fertility (Children per 7 70 Life Expectancy at Birth 6 60 5 50 woman) 4 40 3 30 2 20 1 10 0 0 1950-1955 1955-1960 1960-1965 1965-1970 1970-1975 1975-1980 1980-1985 1985-1990 1990-1995 1995-2000 2000-2005 1950-1955 1955-1960 1960-1965 1965-1970 1970-1975 1975-1980 1980-1985 1985-1990 1990-1995 1995-2000 2000-2005 2005-2010 Source: UN Population Projections, 2010 revision. 15. Pension system dependency ratios improved from 1995 to 2009 while coverage fell then rebounded. As the growth in contributors outstripped that of beneficiaries, the system dependency ratio decreased from almost 55% in 1995 to just over 35% in 2009 (Figure 2). SIGO data suggest that participation in the social insurance scheme was about 50% of the economically active population at end-2005 having declined during the early 2000s and rebounded to about the same level as in 1995. -5- Figure 2. Pension Contributors, Beneficiaries and Economically Active Population (Individuals) 3,000,000 2,500,000 2,000,000 1,500,000 1,000,000 500,000 - 1995 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Contributors Contributors to Health Insurance Pension Beneficiaries Population Economicallly Active Population Source: SIGO Statistical Yearbook, 2009. 16. Projections of pension insurance contributors and beneficiaries suggest that future increases in population old-age dependency ratios will translate into worsening system dependency ratios. Old age dependency ratios are projected to increase from about 11% in 2010 to 53% in 2050 (Figure 3). Similarly, those over age 60 will increase from 7.7% of the population in 2010 to 28.6% in 2050. Not only will the pension insurance system face profound challenges from such an increase, so too will households have to devote an increasing proportion of their income and savings to supporting the elderly. In 2010, life expectancy at retirement was 15.6 for men at age 60 and 22.6 for women at age 55, both of which will increase by 2050 by 3.6 and 4.5 years, respectively (20% and 23% increases - Figure 4). Population aging suggests that the working age population will decline as a proportion of the total population thereby suggesting the need for working longer and retiring later (Figure 3). Similar trends can be observed through projected population pyramids (Figure 5). -6- Figure 3. Age Structure of the Population – 2010-2050 100.0% 100.0% 7.7% 9.3% 11.5% 13.9% 16.6% 90.0% 19.5% 22.6% 90.0% 26.0% 28.6% 80.0% 80.0% 70.0% 70.0% 60.0% 60.0% Ret Age + 67.1% 66.3% 64.1% 62.7% 61.9% 50.0% 61.2% 59.6% 50.0% Age 15-Ret Age 56.5% 53.6% 40.0% 40.0% Age 0-14 30.0% 30.0% Old Age Dep Ratio 20.0% 20.0% 10.0% 25.2% 24.5% 24.5% 23.4% 21.4% 10.0% 19.3% 17.8% 17.5% 17.8% .0% .0% 2010 2015 2020 2025 2030 2035 2040 2045 2050 Source: UN Population Projections, 2010 revision. Figure 4. Projected Life Expectancy and Normal Retirement Age 35.0 30.0 25.0 20.0 15.0 10.0 5.0 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 men (60) women (55) Source: World Bank PROST projections based on data from the UN Population Projections, 2010 revision. -7- Figure 5. Population Pyramids – 2008-2058 (2008) (2058) 2008 2058 97 97 89 89 81 81 73 73 65 65 57 57 49 49 41 41 33 33 25 25 17 17 9 9 1 1 35 25 15 5 5 15 25 35 25 15 5 5 15 25 Source: UN/World Bank estimates. Note: Males are on the left and females on the right of each diagram; the vertical axis is age cohorts. 17. With a growing elderly population and a flat working age population, it is incumbent to consider changes to the current pension arrangements. A fourfold increase in the old-age dependency ratio projected over the next 40 years requires serious reform measures in order to ensure that (i) benefits are adequate; (ii) workers, employers and the Treasury can afford such benefit promises; and (iii) benefit promises can be sustained over time in the face of such dramatic change. C. The Economic Context 18. Recent structural shifts in the Mongolian economy have highlighted the need to revisit the design and implementation of the system of social insurance. These shifts include the increasing importance of mining in the economy, the growing size of the urban informal economy, and the continued concern over old-age income security for herders which are largely outside the current system of social insurance. 19. Growth in the importance of mining and other natural resources in the Mongolian economy along with an increasing importance of private sector employment require flexible social insurance mechanisms to accommodate job mobility and labor market volatility. The Mongolian economy’s rapid recent growth has been fueled by booming mineral exports driven by global prices and rising coal production, ongoing development of large mining projects, surging credit growth, improved agricultural output (primarily a base effect from last year’s severe winter), and expansionary macroeconomic policies. 8 Mining output is set to increase substantially as two large projects underway soon start production. Yet with such rapid growth and dependency on commodity production, Mongolia’s economy will be exposed to greater levels of economic volatility as a result of commodity price fluctuations. Such volatility will 8 See IMF, Public Information Notice on Second Post –Program Monitoring, Public Information Notice (PIN) No. 11/146, November 28, 2011. -8- impact both the labor force and retirees. Volatility in household incomes can be dampened by savings mechanisms and other measures to strengthen the social insurance system. Pensions and other social insurance need to accommodate such anticipated volatility and create incentives for workers to adapt to it. 20. Relatively high historical social insurance contribution rates have contributed to urban job informality and limits in coverage. Relatively high historical contribution rates of 26-28% of wages for social insurance also contribute to the incentives for informality. Generous minimum pension benefits and incentives for early retirement, along with a culture of pooling and sharing resources within the households appear to explain some of the observed job idleness in Mongolia. Social insurance programs need to be designed in a way that encourage work or at least do not discourage it. 9 21. Since the early economic transition in 1990-1993, herders and informal sector workers have effectively been excluded from coverage of pensions and other social insurance. Although many herders since 1993 have had met the full or partial vesting requirements for pensions, this trend is changing with the passage of time. It is therefore important to evaluate mechanisms and incentives to include these so far excluded populations. D. A Growing Fiscal Burden 22. The difference between contributions to the pensions insurance fund and benefits has grown from a fraction of one percent of GDP in 2003 to almost three percent of GDP budgeted for 2010 (Figure 6, Table 3). Total social insurance expenditures have increased substantially from about 5.0 percent of GDP in 2001 to about 6.2 percent of GDP in 2009. Since most social insurance benefits are long-term such as pensions, disability and survivorship; this means that the Government subsidy of the social insurance system has had to increase substantially since 2001 to offset the shortfall between pay-as-you-go contributions and long-term benefits. Old age, disability, survivorship and military pensions represented about 89 percent of all social insurance expenditures in 2009 (Table 3). Disability pensions were about 0.7 percent of GDP during 2009 or about 11.7 percent of all social insurance expenditures. As a revenue source, security contributions have been stable over the past decade at about 4.8 percent of GDP in spite of the 2008 decrease in contribution rates. The data suggests that the reduction in the contribution rate may have been offset by an increase in coverage and real increases in wages. 9 See World Bank, Mongolia: Building Skills for the New Economy, 2007. The report examines the challenges in the labor market in Mongolia as the economy has undergone the transformation into a market economy identifying three major interrelated challenges—joblessness, informality and skills mismatch. -9- Figure 6. Social Insurance Contributions and Benefits (% of GDP) 8.0% Contribution 7.0% income 6.0% 5.0% % of GDP Transfers 4.0% from State 3.0% budget 2.0% Social 1.0% Insurance 0.0% Fund Expenditure 2003 2004 2005 2006 2007 2008 2009 Source: SIGO Statistical Yearbook of Social Insurance, 2010. Table 3: Social Insurance Benefit Expenditures (in percent) Percent of GDP Percent of Total 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2000 2005 2009 Total Pension and benefit 4.9% 5.0% 5.2% 5.3% 5.8% 5.1% 4.6% 5.5% 5.8% 6.1% 100.0% 100.0% 100.0% Old age pension 3.3% 3.3% 3.3% 3.3% 3.7% 3.2% 3.0% 3.6% 3.8% 4.0% 66.2% 64.0% 65.3% Disability pension 0.5% 0.5% 0.5% 0.6% 0.7% 0.6% 0.6% 0.7% 0.7% 0.7% 9.5% 12.7% 11.9% Survivors' pension 0.5% 0.5% 0.6% 0.5% 0.6% 0.5% 0.4% 0.5% 0.5% 0.4% 10.8% 9.9% 7.0% Military pension 0.3% 0.3% 0.3% 0.3% 0.3% 0.3% 0.3% 0.3% 0.4% 0.4% 6.2% 6.0% 6.0% Sickness benefits 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.0% 0.0% 0.1% 0.1% 1.3% 1.2% 1.1% Funeral benefit 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.0% 0.0% 0.0% 1.4% 1.4% 0.6% Pregnancy and maternity benefit 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.2% 1.2% 1.2% 4.0% Work accident & occ. Disease 0.2% 0.2% 0.2% 0.2% 0.2% 0.2% 0.2% 0.2% 0.2% 0.2% 3.3% 3.5% 3.9% Memo: Social Welfare Expenses 0.0% 1.1% 1.2% 1.2% 1.1% 1.6% 2.0% 2.0% 0.0% 0.0% Source: Ministry of Finance and National Statistical Office. 23. Baseline projections suggest that the current scheme will continue to require substantial State subsidies even though retirees under the NDC scheme will begin to retire in 2015 with lower benefits (Figure 7). Baseline projections suggest that in the short-run, as GDP increases substantially and benefits lag GDP growth, the current deficit for the pension system will be reduced somewhat. However, once GDP growth stabilizes to a more modest level and as projected wages increase, the net subsidies required for pension insurance grow from about 2.5% of GDP in 2015 to almost 6.0% of GDP in 2055. Much of the projected negative current balances (and thus required State subsidy) reflect the projected costs of the minimum pension guarantee under the NDC scheme for post- 1960 cohorts, along with a growing absolute size of covered retirees. -10- Figure 7. Projected Baseline Defined Benefit and NDC Current Balances (% of GDP) 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 2011 (1.0%) (2.0%) (3.0%) (4.0%) (5.0%) (6.0%) (7.0%) Source: World Bank PROST projections. E. Pension Insurance Scheme Weaknesses and Reform Needs 24. The defined-benefit scheme applied to cohorts born before 1960 suffers several weaknesses: 10 • The benefit promise when combined with demographic characteristics and the 14% contribution rate results in a need for State subsidies of over 3% of GDP per year; • The retirement age of 60 for men and 55 for women is too low to support a benefit which can moderate the required State subsidy; • Early retirement provisions for workers in hazardous professions, special professions and mothers of four children contribute to the fiscal cost and weak work incentives; • The accrual rate at 2.25% for the first 20 years and 1.5% thereafter creates strong incentives for short work histories, particularly after satisfying the vesting requirement; • The 5 year average for determining the reference wage for benefits creates a number of potential inequities between cohorts; 11 • The scheme has no automatic indexation and historically had years of under-indexation of benefits after retirement which reduces their real value. As a result, many individuals who were not near the minimum pension at retirement soon found themselves subject to the minimum years later; 12 and 10 For details of all the pension system parameters see Appendix 1. 11 For example, when the 5 year average is taken from a period of rapid inflation and/or wage growth (such as in the early 1990s), the entry replacement rate was particularly low. 12 Although current legislation suggests that benefit adjustments should be in line with inflation, actual indexation has in many cases resulted in a depreciation of the real value of the benefit during lengthy periods and real increases in pension benefits during other periods. This creates uncertainty for pensioners and is fundamentally at odds with the objective of providing old-age income security. -11- • Substantial real growth in the minimum wage in recent years has resulted in an estimated 50% of pension insurance beneficiaries receiving the minimum pension of 75% of the minimum wage. The minimum pension guarantee creates strong incentives to satisfy the 20 year vesting requirement though the incremental benefit provided for service after 20 years is reduced (See Table 4). Table 4: Average Wage, Minimum Wage and Minimum Living Standard (Tugriks per month, percent) 2003 2004 2005 2006 2007 2008 2009 2010 Average Wages and Salaries 81,500 93,100 101,200 127,700 173,000 274,200 300,500 330,000 Legal Minimum Wage 25,000 30,000 42,500 53,600 80,000 90,000 108,000 108,000 Minimum Living Standard (Ulaanbataar) 25,300 26,500 30,000 42,800 60,100 73,300 101,100 101,600 Maximum Insurable Earnings 300,000 400,000 425,000 530,000 800,000 900,000 1,080,000 1,080,000 Percent of Average Wages and Salaries Legal Minimum Wage 30.7% 32.2% 42.0% 42.0% 46.2% 32.8% 35.9% 32.7% Minimum Living Standard (Ulaanbataar) 31.0% 28.5% 29.6% 33.5% 34.7% 26.7% 33.6% 30.8% Maximum Insurable Earnings 368% 430% 420% 415% 462% 328% 359% 327% Annual Change Average Wages and Salaries 14.3% 14.2% 8.7% 26.2% 35.5% 58.5% 9.6% 9.8% Legal Minimum Wage 20.0% 41.7% 26.1% 49.3% 12.5% 20.0% Minimum Living Standard (Ulaanbataar) 4.7% 13.2% 42.7% 40.4% 22.0% 37.9% 0.5% Maximum Insurable Earnings 33.3% 6.3% 24.7% 50.9% 12.5% 20.0% Average Annual Inflation 5.1% 8.2% 12.7% 5.1% 9.0% 25.1% 6.3% Source: National Statistical Office, Statistical Yearbook 2008, MLSW, SSIGO Statistics 2010, World Economic Indicators, 2011. 25. The abrupt reduction in benefits for those under the NDC scheme born in 1960 and lack of a transition mechanism creates highly inequitable levels of income replacement between cohorts. Projections suggest that replacement rates for new retirees will fall substantially from 40-50% for pre-1960 cohorts to about 32-35% for post- 1960 cohorts (Figure 8). 13 The figure also indicates the substantial importance in the NDC benefit formula of the minimum pension guarantee. 13 Workers in the defined benefit scheme with 10-19 years of service receive a lower minimum benefit amount while workers with less than 10 years of service receive a return of contributions. Benefits under the NDC scheme for workers born after 1960 are determined based on the contribution rate credited to an individual account, the rate of return applied to the notional balance (the three-year rolling average rate of growth in average reported wages), life expectancy at retirement using unisex retirement tables and the minimum pension provision. -12- Figure 8. Baseline Defined Benefit and NDC Replacement Rates 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% .0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 pre-1960 cohorts post-1960 cohorts-NDC post-1960 cohorts-min pension top-up Source: Bank estimates using the PROST model. Note: Replacement of average covered wages by new male retirees. The dark shaded area under the NDC curve indicates the benefit provided according to the NDC formula while the lighter shading is the benefit attributed to the minimum pension guarantee. 26. The NDC minimum pension guarantee will create a substantial benefit floor at a considerable cost. • The wage distribution and contribution histories will result in most retirees entitled to the minimum. This creates weak incentives for workers with short contribution histories and/or low wage bases for contributions. It also creates a mismatch between cumulative contributions and benefits requiring a state subsidy. 14 Since the average service period for recent retirees is about 25 years, this would result in a minimum pension equal to about 25 percent of the average covered wage. 15 Many workers therefore have an incentive to under-report wages in anticipation of receiving the relatively generous minimum pension which is computed based on the average covered wage. • The financing formula for the minimum pension (a portion of 4 percent of contributions or 0.56% of covered wages) 16 will be insufficient to finance the anticipated costs of the difference between the old-age pension generated by the notional account accumulation and the minimum pension guarantee in the future. The minimum pension therefore either requires continued fiscal subsidies or would need to be modified or replaced by an alternative instrument. 14 PROST baseline projections suggest that over the projection period an estimated 85 percent of the recipients of old-age pensions under the NDC scheme will receive the minimum pension. 15 The formula for the minimum pension guarantee is 20% of the average covered wage after 15 years of contributions plus 0.5% of the average wage for each additional year of contributions. 16 The 4% is also supposed to cover the administration cost, and additional costs arising from disability and survivor benefits. -13- • The generosity of the NDC scheme minimum pensions guarantee creates segmentation in the labor market, whereby those workers who reach the minimum vesting requirement receive a relatively generous benefit while those not meeting the vesting requirement receive nothing other than the Social Welfare Pension benefit. 27. The retirement ages for regular retirees (aged 55 for women and 60 for men) and those in special occupations (as low as age 45) are too low to support a meaningful replacement rate at retirement in the NDC scheme. With the average age of retirement for women being about 52, such workers would, on average spend 27 years working and 23 years in retirement. Mongolia’s retirement ages are also lower than most comparator countries in Eastern Europe and Central Asia (Figure 9). The prevalence of workers under the NDC scheme that are projected to receive the minimum pension is also symptomatic of the low pension benefit level which to a large degree results from the retirement age being low relative to the life expectancy at retirement. Figure 9. Retirement Age for Pension Insurance Schemes in East Asia, Europe and Central Asia 70 60 50 Retirement Age 40 30 20 10 0 Men Women (or Men and Women if the Same for Both) Source: World Bank database. 28. There is a tradeoff between the adequacy of pensions and other social insurance benefits on the one hand, and on the affordability of the sustainable contribution rate on the other. The combined contribution rate includes contributions to short-term, work injury, unemployment and health (Table 5). Under the NDC scheme, an estimated 13.44 percent of covered wages is applied toward an individual’s account and the remaining 0.56 -14- percent is to be used to fund the costs of administering the system, provisions for survivor and disability pensions, and provision of a minimum pension guaranty. 17 Table 5: Social Insurance Contribution Rates (in percent of covered wages) Self-Employed Benefit Employer Employee Total (voluntary Pensions 7.0 7.0 14.0 10.0 Short-term benefits 0.5 0.5 1.0 1.0 Work injury benefits 1.0-3.0 1.0-3.0 1.0 Unemployment Benefits 0.5 0.5 1.0 MNT 5360/month* Health Insurance 2.0 2.0 4.0 MNT 670/month** Total 11.0-13.0 10.0 21.0-23.0 *Foreigners and stateless persons ** herders, students, persons whose contributions are paid by the Government, prisoners Source: SIGO. 29. Although the 21-23 percent contribution rate creates incentives for evasion and wage underreporting, the rate is also lower than for most of the former transition economies in Europe and Central Asia (Figure 10). Several transition economies such as Mongolia are at least partially financing the legacy costs of former universal coverage through pay-as-you-go financing in the pension system and therefore have had to maintain relatively high contribution rates in order to finance the benefit entitlements of their retired populations. The contribution rate is modest when compared to other non- transition countries in East Asia however which tend to be newer schemes with lower benefit promises and little if any legacy costs. 17 The 1999 legislation specified that 4 percent of wages would be used to finance disability and survivorship benefits and administrative costs. The 2008 legislation which decreased total pension insurance contribution rates from 19% to 14% did not specify how much should be credited to individual notional accounts. The 0.56% rate represents .04 * .14 which is our understanding of the appropriate administrative costs based on discussions with officials from SIGO. However, our review of administrative data suggests that the actual amounts credited to notional individual accounts were lower than this. It is important to clarify what individual entitlements are and to publicize such entitlements to ensure transparency and accountability. -15- Figure 10. Comparative Contribution Rates for Pension Systems in East Asia, Europe and Central Asia (% of covered wages) 40.0 35.0 Contribution Rate (%) 30.0 25.0 20.0 15.0 10.0 5.0 0.0 Philippines Turkey Latvia Serbia Moldova Lithuania Azerbaijan Bulgaria Armenia Korea, Rep. Poland Hungary Georgia Belarus China Albania Indonesia Lao PDR Thailand Mongolia Vietnam Bosnia and Herzegovina Romania Ukraine Malaysia Russian Federation Kyrgyz Republic Total Employer Employee Source: World Bank database. 30. The 1999 legislation reforming the Pension Insurance System includes inconsistencies between the NDC benefit formula and the financing plan. The Government envisioned the introduction of the NDC scheme as an interim step towards moving to a partially-funded defined-contribution scheme in the context of largely undeveloped capital markets and limited fiscal resources. However, the benefit formula in the legislation is unaffected by the rate of return of funds under active management and therefore creates conflicting signals and potential confusion. Since 1999, disbursements to retirees under the defined-benefit scheme have exceeded contributions and in 2008 a reduction in the contribution rate increased the annual deficit. Thus no reserves have been actively managed, nor is it likely that movement from notional to partially-funded accounts will happen because continued net deficits are projected for the future. 31. Disability benefits under the NDC scheme suffer from several weaknesses including no articulated link to contributions. A worker losing more than 50% of capacity is entitled to an annuitized benefit for life equal to 60% of the average of the last three years’ wages, regardless of age and the notional balance in his or her account. This is a benefit with a higher replacement rate than has been projected for even a full term worker with a 40 year work history receiving an old-age benefit. Benefits for the disabled are thus significantly greater than for those receiving an old-age benefit. High annuitized (defined) disability benefits without re-certification of the disability create pressure on the medical examiner to certify individuals as disabled in an effort to provide assistance in the face of inevitable volatility and transitions in the economy such as economic crises, -16- bankruptcy, unemployment, etc. In addition, because there is no actuarial reduction for the benefit based on age, there are strong incentives for people prior to retirement age for certification as disabled in order to extend the period of their service retirement. 32. There is also room for improving the survivorship benefit formula under the NDC scheme. Under the NDC scheme, survivors receive a benefit of between 40% and 60% of the monthly average wage in the last three years. 18 The benefit provides a minimum replacement rate for survivors far higher than for old age. Moreover, the benefit is not linked to a dedicated contribution or subject to long-term actuarial projections ensuring that contributions can support the benefit formula. 33. Some allowances are excluded from the wage base for determining both pensions contributions and benefits. Although this has the effect of lowering contributions when measured against total emoluments, it also has the effect of reducing the effective replacement rate when also measured in respect of total emoluments. Moreover, different workers may receive different proportions of their compensation as allowances inviting issues of equity of treatment between workers. 19 Thus the inclusion of all components of compensation in the calculation of contributions and benefits substantially improves benefit adequacy albeit comes at the expense of affordability of contributions. Care is needed in devising a transition to include non-pensionable allowances for older cohorts as such inclusion could increase the fiscal burden for the defined-benefit scheme. In the medium-term, the Government could review the scope of compensation subject to social insurance contributions in conjunction with a similar review of treatment for corporate and personal income taxes. 34. The practice of exempting contributions, accumulations and benefit distributions from personal income taxes is too generous. Although it creates a tax benefit which favors compliance in the mandatory pension scheme, it also results in forgone fiscal revenues. The tax benefit is also greatest for those with the highest incomes. If contributions and accumulations are exempt from tax, then distributions should be subject to personal income tax when received, consistent with the personal tax rate and thresholds applicable. Such consumption tax would provide for a more equitable outcome. 35. Overall, the changing needs of Mongolia’s economy and aging society, the projected fiscal burden of almost 6% of GDP per year over the long term, and the technical design weaknesses indicated, together provide a strong rational for reforms 18 A spousal survivor would receive such a benefit from the date of death up to retirement age. Young children receive survivorship benefits up to age 16 or if students age 19, or if incapacitated for the period of loss of working capacities. 19 The replacement rate is lower if calculated as a proportion of total pre-retirement income compensation including allowances. Many countries are moving to include allowances in the wage base. -17- to the Pension Insurance scheme, Social Welfare Pension and voluntary arrangements as suggested in the section below. -18- III. Objectives and Options for a Design Architecture A. Objectives and Design Considerations 36. The core objectives of most systems of social insurance and social assistance to ensure old age income and poverty protection are:20 (i) To prevent a material drop in consumption from work life into retirement when one may be less able to work; and (ii) To protect the elderly against poverty in old age or extreme old age when they may be unable to work. Such systems may also redistribute resources within or between generations. Social insurance systems also often insure against the risks of death or disability of a wage earner. 37. In seeking to achieve such core objectives, policymakers should weigh the tradeoffs between (Figure 11): (i) Adequacy of the benefit; 21 (ii) Fiscal, employer and employee affordability and sustainability; 22 and (iii) Retirement age 23 and related qualifying conditions for receiving a benefit. These are the fundamental design choices for consideration by policymakers. 20 See World Bank (2008), The World Bank Pensions Conceptual Framework. 21 Adequacy is often evaluated based on the replacement rate and elderly coverage. Replacement rate refers to the pension benefit as a proportion of the individual pre-retirement wage and as a proportion of the average covered wage. Coverage for the purposes of evaluating adequacy is often considered as the proportion of individuals receiving old age benefits as a proportion of elderly who have reached the retirement age. 22 Employer and employee affordability is often evaluated based on the contribution rates and wage base subject to contributions. Fiscal affordability refers to the burden of fiscal subsidies required over time to ensure that pension promises are fully met. Indicators of fiscal affordability include required state subsidies as a proportion of GDP or as a proportion of projected fiscal revenues. 23 This refers to the age at which pension system contributors are able to receive benefits. Actual retirement from the labor force often is later than the age at which benefits can be received. -19- Figure 11. Tradeoffs in Achieving Core Objectives of Old Age Income Protection Affordability 14% Employer/ Employer/Worker/ Worker Contribution Government Rate and est. 20.1% Contribution (Contribution Rate Rate from State and Fiscal Costs) Subsidy Adequacy Work- (Target Retirement Replacement Life expectancy at Balance 25-50% Rate) Replacement Rate? age 60 (2010) 17.6 (Retirement years (men), 21.9 (Man working 25 Age) years (women). years) 38. All pension designs have to weigh the tradeoffs between parameters to achieve desired levels of adequacy, affordability, sustainability and work/retirement balance. In this way, whether the authorities choose to continue with the NDC design for post-1960 cohorts, adopt a points design or establish a sustainable defined-benefit scheme, in all cases they will need to weigh these variables in establishing the key parameters of the schemes. This report therefore suggest that policymakers first consider the overall objectives of income replacement, elderly poverty protection as well as tradeoffs between affordability, benefit adequacy and work/retirement balance prior to considering architectural designs to achieve these objectives. 39. A multi-pillared design can provide a diversified set of instruments to accommodate the needs and risk characteristics of different population groups (Figure 12). 24 In this way, a multi-pillar design can strengthen the incentives for old age income and poverty protection. Each pillar performs a separate function while together they comprise a flexible system for old age income protection. The Tier I - Basic Pillar ensures minimum income support for all elderly; the Tier II – Contributory Pillar provides a savings and social insurance vehicle for an annuitized benefit in old age; while Tier III – the Supplementary Pillar provides a supplementary savings arrangement which can be provided by government, enterprises, or organizations to their workers as a means of supplementing income from the Tier II Contributory Pillar. Tier III arrangements can also 24 Such a multi-pillar approach has also been endorsed by the Ministry of Labor and Social Welfare. -20- afford the opportunity for individuals, the self-employed and others not covered by Tier II social insurance to save for their long-term pension needs. Figure 12. A Three Tier Approach to Old Age Income Protection Function Objective Instrument Basic Pillar – Protection against Targeted Social Pension – Non- Tier I elderly poverty contributory targeted benefit Prevention of a material drop in Contributory Mandatory Contributory consumption due to Pillar – Tier II Pension Insurance old age, disability or death Supplementary Income supplements Supplemental occupational and Pillar –Tier III beyond pension individual savings arrangements insurance system 40. Each pillar has distinct characteristics and exposes stakeholder to different types of risks. 25 So, for example a Tier I targeted social pension may ensure that the elderly have a minimum level of income support while at the same time exposing them to inflation or political risks. Similarly, a Tier II benefit can be designed to cover longevity, other demographic and investment risks for contributors but exposes the sponsor (in this case the Government) to the same risks. Tier III risks also depend upon the design. Generally the contributors are subject to the portfolio investment risks, regulatory risks for the custodian, data and asset managers, and risks that contributions may not be fully remitted to a financial intermediary. There can also be risks during the payout phase. As many of the risks between the different tiers are not fully correlated, the mix of the different tiers can diversify the risks which are borne by participants. 41. Together, the proposed combination of tiers or pillars can strengthen old age income protection for Mongolians. If well designed, the parameters can also mitigate the fiscal exposure to support old age income protection by targeting public support to lower income beneficiaries. Finally, a well-considered design can improve the incentives for participation and compliance. 25 See Robert Holzmann and Richard Hinz, Old Age Income Support in the 21st Century, World Bank, 2006. -21- B. Tier I – Targeted Social Pensions B1. Rationale for a Targeted Social Pension 42. A targeted social pension is one means of ensuring close to universal elderly coverage by some form of pension provision and therefore overall pension adequacy. Elderly coverage by social insurance is projected to decrease over time as those retirees enjoying vested benefits in part from service prior to 1994 gradually diminish in number. As the pension insurance scheme relies on wage-based payroll taxes for contributions, it only covers about half of the working age population and such coverage is unlikely to increase substantially in the near future. A targeted social pension can ensure elderly coverage as beneficiaries under the pension insurance scheme are gradually reduced. 43. A targeted social pension can also be designed to supplement a contributory (Tier II) pension scheme to reduce the coverage gaps created by vesting requirements and the lack of integration between the current Social Welfare Pension and the social insurance pension. 44. The pension insurance scheme also cannot fully accommodate the special needs of herders, informal workers, and formal workers with insufficient lifetime contributions to support an adequate pension. A majority of workers in Mongolia are informal and of these workers, most are low skilled self-employed workers in agriculture (mainly herders), who lost the universal ‘right’ to a pension in 1995. 26 The increased informality among salaried workers (19 percent in 2006) is partly explained by social security taxes which induce firms to hire workers without contracts and overuse (and misuse) temporary contracts (World Bank 2007), which carry no social security benefits. There is considerable concern amongst the authorities over the limited coverage of the contributory schemes, yet options for improving such coverage have not yet been fully explored. 45. Finally, a well-designed social pension can impact income redistribution by targeting lower-income retirees. A means-testing targeting mechanism and/or pensions test can be designed so that those retirees with the least pension insurance benefits receive the largest social pensions. 46. The objectives in establishing a Social Pension under consideration evidently are to: (i) ensure a minimum benefit level to all elderly regardless of contribution histories in the pension insurance scheme; (ii) to reduce or eliminate the elderly population that has 26 Informal workers are defined as those in low productivity jobs with no social security benefits. See World Bank (2007). -22- little or no other forms of income or assistance; and (iii) when combined with the Tier II component, to effect a higher level of income replacement for lower income retirees. B2. Recommendation, simulated benefits and fiscal costs 47. A targeted social pension is suggested with the characteristics below (Figure 13): • Objective. Basic absolute income support for the bottom half of the retiree income distribution is suggested. • Targeting. A categorical benefit targeting the elderly is justified in view of institutional limitations and the proposed integration with the proposed Tier II pension insurance scheme. • Means-testing. A pensions test is proposed with those retirees receiving at or above the average level pension insurance benefit only entitled to a nominal amount of a benefit. 27 • Benefit form, level and factor reductions. The benefit form should be an absolute amount. The maximum benefit should be determined to target an elderly poverty gap. Sixty percent of the MLS for Ulan Bataar has been proposed as an initial benefit amount though further analysis is needed to calibrate the precise level. A pensions test - factor reduction has been suggested whereby retirees will have their targeted social pension benefit amount reduced by a proportion of their pension insurance benefit. • Age and other qualifying conditions. A minimum age for qualification of 60 (men and women) has been suggested and this age should be gradually increased to age 65 in line with increases in the retirement age for the pension insurance scheme. • Indexation provisions. All benefits are proposed to be linked to the MLS. This has the effect of ensuring that new and old retirees receive the same benefit and also implicitly indexes the benefit to the change in the MLS. In the medium-term, an elderly basket of minimum living requirements should be developed which can serve as the basis for benefit determination and indexation. 27 A benefit of about MNT 3,000 per month ensures that all recipients receive something yet limits the cost of such benefit. -23- Figure 13. Summary of Design Considerations for Tier I Targeted Social Pensions Issue Policy Options Recommendation Objective • The benefit may target an absolute level of income such as the gap The benefit should seek to achieve an between the individual’s level of income and the individual poverty absolute level of income such as a line or the MLS; or proportion of the individual MLS for • The benefit may target the gap between individual income and a Ulan Bataar but that consideration is relative income level such as a proportion of average covered wages. given to gradually move to a relative Many OECD countries set the benefit level based to achieve a income level approach. minimum relative retiree income level. General • Many countries have little or no elderly social assistance which A social pension which targets lower social targets individuals based on age but, rather, utilize social assistance to income elderly can be justified based on assist- poor households including those with elderly. In cases where a the rationale that the household means- ance vs. country has in place a well-established targeting mechanism, there testing arrangements are limited. social often is a weak policy rationale for a targeted elderly social assistance Moreover, a minimum benefit with an pension program because the elderly often are in households which are better objective of elderly poverty protection off that those without elderly. Currently there are institutional is being considered in parallel with limitations in the targeting mechanisms for social assistance in social assistance to poor households. Mongolia. • Social assistance which targets the elderly can also be supported based on a rationale behind the policy and/or institutional integration with other instruments. Means • Means-testing criteria can be applied using ex-ante income, No ex-ante means-testing criteria to testing consumption or proxy-means testing criteria. determine eligibility is suggested but • Targeting effectiveness depends upon several variables including the application of a “pensions-test� to targeting objective, threshold applied and institutional capacity. reduce or eliminate benefits for • Means testing often impacts economic efficiency whereby individuals individuals receiving pension insurance will adjust their savings, work and consumption patterns in order to benefits over a threshold to be qualify for a benefit. determined. This threshold could be the average pension insurance benefit level. Benefit • Benefits can be designed as a “top-up� where the level is sufficient to • A benefit level of perhaps 60% of the design, close an identified individual poverty gap bearing in mind the MLS could be considered though further level and difficulty in identifying incomes of the elderly and the tendency for analysis of distributional and cost reductions misreporting. implications is needed. 28 or claw- • Benefits can be an absolute amount (such as a % of the MLS) targeting • A benefit factor reduction is proposed backs an elderly poverty gap which then ignores the distributional impact of through the application of a pensions individuals with different income or consumption patterns under the test. The factor reduction should be a eligibility threshold. proportion of the pension insurance benefit for retirees receiving at or below • Benefits can be partially reduced for existing pension income from the average benefit; little or no social social insurance. This is administratively the easiest to achieve and, if pension benefits should be provided to well designed, can result in minimizing the behavioral effects. those receiving more than the average. Age & • The age of entitlement may or may not be linked to the age of receipt A minimum age of 60 for qualification other of a pension insurance benefit. Some countries mitigate the adverse (men and women) is suggested as per qualifying incentives to participate in the labor market (and constrain costs) by the current social welfare pension conditions having a higher age of entitlement for a social pension than for a phased to increase to age 65 as and if social insurance pension. the pension insurance retirement age is increased. Indexation • Benefit indexation is strongly linked to the initial benefit level. • A benefit level of perhaps 60% of the • Benefit indexation can be implicit or explicit; and linked to the MLS MLS is suggested which is then basket, consumer price inflation, covered wages or a mix thereof. If implicitly indexed to the growth in the benefits are determined as a proportion of wages but indexation MLS. linked to prices then new beneficiaries will get higher benefits than • Ideally the indexation should be old ones. Since wages grow faster than prices over time, there are linked to an elderly consumption also substantial differences in simulated costs between a benefit basket in the medium-term. 28 An analysis of the real increase in the MLS since 2004 and the relationship between the MLS and the average covered wage raises a number of concerns about the MLS calculation methodology. In May, 2011, the MLS represented about 40% of the average covered wage. Although 60% of the MLS has been suggested as a social pension benefit level, this is meant only as an initial suggestion. Additional analysis is needed of the MLS methodology (including a review of the consumption basket) and evaluation of different benefit levels of the social pension against distributional data in the 2008 Household Income Survey. In approving the most recent revision to the MLS, the Parliament evidently also requested a review of the methodology for its calculation. The proposed benefit level of 60% of the MLS is an effort to balance what may be an inflated MLS level which could be revealed by the review of the methodology. -24- indexed to wages and one indexed to prices. 48. There is a tension between encouraging savings for retirement in the mandatory pension insurance or voluntary savings schemes proposed on the one hand, and, on the other hand, providing minimum elderly income support through a social pension. A social pension may create incentives to reduce work, savings and intra-family transfers. Severe means testing criteria runs the risk of large errors of exclusion. Further, low benefit levels may not achieve the elderly protection desired. 49. The proposed factor reduction is designed to reduce the social pension benefit based on contributory pension income while ensuring a positive marginal additional benefit for contributions to an individual’s notional individual account. This is illustrated in Table 6 below. The targeted social pension benefit is adjusted by a reduction factor which gradually increases up to a level of 95% when the Tier II benefit is greater than 42% of the average wage. 29 This design aims to create a positive incentive to continue to contribute to the pension insurance scheme for those with limited years of service and/or low salary levels. Table 6: Stylized Illustration of Factor Reduction of Targeted Social Pension Individual Tier II Pension as a % of Average Wage 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% Factor Reduction 0% 15% 40% 65% 90% 95% 95% Tier I Social Pension/Month Before Factor Reduction (Tugrits/Month - 60% of MLS) 60,960 60,960 60,960 60,960 60,960 60,960 60,960 Tier I Targeted Social Pension/Month After Factor Reduction (Tugrits/Month) 60,960 51,816 36,576 21,336 6,096 3,048 3,048 Tier II Contributory Social Insurance Pension Income (Tugrits/Month) - 33,000 66,000 99,000 132,000 165,000 198,000 Total Tier I and Tier II Income after Factor Reduction (Tugrits/Month) 60,960 84,816 102,576 120,336 138,096 168,048 201,048 Source: Staff estimates. Note: based on an assumption that the amount of the Targeted Social Pension would be 60% of the Minimum Living Standard. 50. The projected fiscal costs of a Targeted Social Pension largely depend upon the benefit level, factor reduction and indexation (Figure 14). 30 The costs substantially increase in line with the growth in the size of the population over age 60 or 65. Much of this cost increase can be offset (when measured as a proportion of GDP) if price indexation is chosen because of the difference between price growth on the one hand, and GDP growth on the other. If the benefit was to be indexed to the change in the MLS, this increase would likely be no less than price growth based on the formulation of the MLS. The costs of such indexation could therefore be similar to that for price-based indexation. 29 The rationale behind such a level is that a replacement rate of 42% of the average wage represents an average benefit as a percent of average wage for new retirees in 2011. 30 The behavioral dynamics which would result in a change in the contribution patterns in the pension insurance system have not been modeled as a result of the establishment of a targeted social pension. -25- 51. Figure 14 illustrates the projected fiscal costs of a Targeted Social Pension based on different age and indexation parameters. A benefit targeted at retirees earning less than the average Tier II pension at age 60 with an initial benefit level of 60% of the MLS, price indexation and a factor reduction for a proportion of pension insurance benefits would have a projected fiscal cost of about 0.3% of GDP in the short term rising to about 0.75% of GDP by 2050 before declining to below 0.5% by 2082 (Figure 14). Although this would appear to be fiscally manageable by most metrics, it is useful to compare the projected fiscal costs according to different eligibility age and indexation parameters. Figure 14. Projected Fiscal Costs of a Targeted Social Pension 3.0% 2.5% Age 60 Price 2.0% Indexed 1.5% Age 65 Price Indexed 1.0% Age 65 Swiss Indexation 0.5% Age 60 Swiss Indexation 0.0% 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048 2052 2056 2060 2064 2068 2072 2076 2080 Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. Assumption is an initial benefit of 60% of the MLS. Swiss indexation refers to benefit adjustments based one-half on price increases and one-half on covered wage increases. The calculation assumes a factor reduction to account for other benefits received under the pension insurance system for covered beneficiaries based on the PROST outputs for benefits. 52. There is a tradeoff between fiscal affordability and benefit adequacy as illustrated in Figure 15. Since over the long-term wages tend to rise faster than prices, a social pension that is indexed to prices will decline relative to average wage levels. One approach is to ensure social pension adequacy while at the same time encouraging contributions to the pension insurance scheme or voluntary contributions to a supplementary scheme so that account balances are sufficient for a basic subsistence during retirement. -26- Figure 15. Projected Benefits of a Targeted Social Pension (% of average covered wage) 20% 18% Benefit as a % of average wage 16% 14% 12% 10% 8% 6% 4% 2% 0% 2012 2015 2018 2021 2024 2027 2030 2033 2036 2039 2042 2045 2048 2051 2054 2057 2060 2063 2066 2069 2072 2075 2078 2081 Wage Indexed Price Indexed Swiss Indexed Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. Assumption is an initial benefit of 60% of the MLS. Wage indexation refers to benefits which are adjusted each year according to the growth in the covered wage. Swiss indexation refers to benefit adjustments based one-half on price increases and one-half on covered wage increases. 53. The proposed benefit is progressive and redistributive: The composite benefit between the Tier 1 Targeted Social Pension and Tier II Pension Insurance Scheme is progressive for those receiving a pension insurance benefit less than the average (Figure 16). Those retirees with low or no benefits under the pension insurance scheme would receive substantially higher levels of income replacement in retirement than higher income individuals (Figure 17). The Targeted Social Pension therefore benefits those with the no alternative pensions or the lowest pension insurance benefits. -27- Figure 16. Stylized Illustration of Combined Benefits of Tier I and Tier II Schemes 250,000 200,000 Monthly Benefit (MNT) 150,000 100,000 50,000 - 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 110% 120% 130% 140% 150% Individual Pension Insurance Benefit as a % of the Average Pension Insurance Benefit Social Pension Contributory Pension Source: Bank estimates. Note: Assumes a Tier I Targeted Social Pension benefit of 60% of the Minimum Living Standard or MNT 60,960 at December 31, 2010. Figure 17. Stylized Example of Replacement Rates from Combined Tier I and Tier II Schemes 250% Benefit as a % of Individual Pre-Retirement Wage 200% 150% 100% 50% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Individual Pension Insurance Benefit as a % of the Average Benefit Targeted Social Pension Contributory Pension Source: Staff projections. Note: Assumes an average replacement rate from Tier II Pensions of 42% and same factor reduction scale as in Figure 16 above. The replacement rate for individuals not receiving Pension Insurance benefits cannot be determined and for those with no wages, the replacement rate would be infinite. 54. A well designed Tier I Targeted Social Pension can complement a Tier II Pension Insurance scheme and strengthen old age income protection in Mongolia. At the same time, such a design can address elderly vulnerability and current gaps between covered and uncovered populations. Over time, a Targeted Social Pension will either decline in its adequacy of wage replacement or become increasingly fiscally expensive to maintain. In view of such a tradeoff between affordability and adequacy, the long-term strategy should -28- be to support measures to ensure that the contributory Tier II pension prevents elderly poverty for the vast majority of the population through measures to ensure full coverage and sufficient contributory work histories while at the same time individuals with short work histories or very low wages are provided with a small guaranteed benefit to be kept out of poverty. C. Tier II – Contributory Pension Insurance C1. Pension Insurance for Old Age Retirement C1.1. Modifying Key Parameters in the existing NDC Scheme 55. This section evaluates the impact on benefits and fiscal costs of parametric reforms to the current NDC scheme for post-1960 cohorts to strengthen benefit adequacy, fiscal affordability and work/retirement balance; adoption of a Points Scheme and FDC scheme were also evaluated as requested by the authorities. This report has evaluated the impact of: • Removing the minimum pension guarantee and replacing it with a Tier I Targeted Social Pension; • Gradually increasing the retirement age to 65 for men and women so that workers have a stronger incentive to work longer and bear more of the burden for their retirement income replacement; 31 • Gradually reinstating a 19% contribution rate to improve benefit adequacy; • Adopting an annuity factor which is actuarially fair; and • A combination of these measures. Replacing the NDC scheme with a Points Scheme or Defined Benefit scheme with the same contribution and target income replacement parameters was also assessed. Finally the potential benefits and risks of an FDC scheme were evaluated. 56. The rationale behind increasing the retirement age is that this has the effect of increasing benefit levels (from increased contribution histories and shorter payout periods) as well as strengthening the incentive to work to an older age. It is proposed that the increase be a gradual one, whereby the age is increased 6 months for each year thereby creating an incentive to work longer while not creating a shock for those close to retirement. The impact of a 6 month per year phased increase would be that this increase would take a period of 20 years for women to reach a retirement age of 65 and men a period of 10 years to reach a retirement age of 65. 31 It is worth noting that that the NDC design already has a strong incentive to work longer built into the benefit formula. -29- 57. Workers in hazardous or special professions and women with four or more children should be subject to a more accelerated transition process. The purpose of such an accelerated transition is to align these professions to the terms and conditions of the rest of the participants in the pension insurance scheme. As suggested below, workers in hazardous professions should be entitled to supplementary Tier III benefits subject to separate qualifying conditions but such Tier III provisions should be governed and regulated apart from the pension insurance scheme. 58. The rationale behind increasing the contribution rate is to achieve a higher benefit and reduce the required State Subsidy. Increasing the contribution rate, along with other recommended measures, can therefore facilitate the earmarking of fiscal resources towards a Targeted Social Pension which is aimed at the lower half of the retiree income distribution. It is also recognized that increasing the contribution rate can come at the expense of the incentives for participation, compliance and proper wage reporting. In this case, the authorities should carefully weigh the tradeoffs between adequacy (though the replacement rate), and affordability (through the contribution rate). 59. The technical issue of making the annuity factor actuarially fair is as follows: The current NDC benefit formula provides for an annuity calculated as the notional account balance divided by the life expectancy at retirement age. This may not actuarially fair to retirees because the notional account balances will and should accrue additional notional interest during the payout phase of retirement which is in excess of the discount rate which would be applied to such interest when computing an annuitized benefit. If such additional notional interest were considered, the benefit and income replacement would be increased. Ensuring such an actuarially fair annuity factor increases benefits at the expense of the system finances though in the long-term can result in close to a zero current balance. 60. Actuarial projections suggest that eliminating the minimum pension guarantee (replaced with a targeted social pension), increasing the retirement age to 65, increasing the contribution rate to 19%, and adopting an actuarially fair annuity factor has the effect of increasing replacement rates and eliminating required state subsidies for Tier II pensions (Figure 18 and Figure 19). 32 As suggested, the fiscal impact of eliminating the minimum pension guarantee also eliminates the projected requirement for a state subsidy of about 6% of GDP in the long-term. This elimination has the effect of reducing the replacement rate for men by about 7-9% of average covered wages. This reduction however, would be offset by an increase in the replacement rate by about 8-10% of wages for men which could be realized by increasing retirement ages, contribution rates and changing annuity factors. 33 Such replacement rates would be for Tier II pensions only; 32 The individual effects of the three reform scenarios and the combined reforms is illustrated in Appendix 5. 33 The increase is slightly less for women at about 6-10% of average wages. -30- they would not reflect the increase in income replacement provided by the combination of Tier I and Tier II benefits for those retirees in the bottom half of the retirement income distribution. The result of these reforms would be a pension insurance scheme that provides about a 40-44% replacement of average wages for newly retiring men at age 65 with a 19% contribution rate. Additional increases in replacement rates could be realized by extending the work histories substantially above those observed through measures to achieve increased participation and compliance. The 40-44 percent replacement rate would be supplemented for by Tier I benefits for lower income retirees and for others by Tier III benefits from voluntary savings arrangements to achieve the full package of old age support. These would be in addition to support from individual savings and family support. Figure 18. Reform Scenario (2E): Removal of the Minimum Pension, Increase in Retirement Age, Increase in Contribution Rates and Adoption of Actuarially Fair Annuity Factor (current balance as a % of GDP) 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 (1.0%) (2.0%) (3.0%) (4.0%) (5.0%) (6.0%) (7.0%) baseline reform 2e Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. -31- Figure 19. Reform Scenario (2E): Removal of Minimum Pension, Increase in Contribution Rates and Adoption of Actuarially Fair Annuity Factor (replacement of average covered wages for new male retirees) 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% .0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 pre-1960 cohorts post-1960 cohorts-baseline post-1960 cohorts-reform 2e Source: Bank estimates using the PROST model. Note: Replacement of average covered wages by new male retirees. C1.2. Evaluation of a defined-benefit scheme 61. The fiscal cost and impact on income replacement of establishing a defined- benefit scheme for post-1960 cohorts was evaluated. Simply extending forward the date to which the NDC scheme would apply would result in continuing the same parameters in the current defined-benefit scheme which have resulted in high and growing fiscal costs and weak incentives (See Reform 3A in Figure 20 and Appendix 5, Figure 37). 62. Reinstatement of a defined-benefit scheme for post-1960 cohorts would require additional changes in parameters to achieve the same features as a modified NDC scheme above. These changes would include: (i) gradually increasing the retirement age to 65; (ii) gradually increasing the contribution rate to 19%; (iii) increasing the years of wages used for calculating the benefit over time to a lifetime wage base and indexing or “valorizing� these wages to the average growth of covered wages; and (iv) reducing the accrual rate from the current 2.25% for the first 20 years and 1.5% thereafter to about 1.1% for each year. This accrual rate was calculated as a sustainable rate over the long- term provided the other parameters were in place (See Figure 20 and Appendix 5, Figure 37). 63. The impact on fiscal costs and replacement rates of reinstating the defined- benefit scheme with revised parameters is minimal when compared to the current NDC scheme (Figure 20, Figure 21). The key reasons are that the parameters for such a sustainable defined-benefit scheme (retirement age, contribution rate, accrual rate, minimum pension and indexation) were set to align with the reformed NDC scheme which -32- is also sustainable. There is therefore little difference in terms of costs or benefits between reformed NDC and DB schemes, provided that the parameters are aligned. Figure 20. Reform Scenarios 3A,B,C: Establishment of Defined-Benefit and Points Schemes for Post-1960 Cohorts (current balance as a % of GDP) 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 (1.0%) (2.0%) (3.0%) (4.0%) (5.0%) (6.0%) (7.0%) (8.0%) (9.0%) (10.0%) baseline reform 3a reform 3b/3c Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. Figure 21. Reform Scenarios 3A,B,C: Establishment of Defined-Benefit and Points Schemes for Post-1960 Cohorts (replacement rates) 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% .0% 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 2011 pre-1960 cohorts post-1960 cohorts-baseline post-1960 cohorts-reform 3a post-1960 cohorts-reform 3b/3c Source: Bank estimates using the PROST model. Note: Replacement of average covered wages by new male retirees. 64. An advantage of the defined-benefit scheme is that it may be more easily understood by a number of members. Disadvantages are: (i) adjustments will need to be undertaken periodically in the contribution rate, accrual rate and/or retirement age in order to ensure its sustainability in the face of unanticipated demographic shifts; (ii) an actuarially fair penalty will need to be legislated for anyone who retires prior to the -33- retirement age and an actuarially fair benefit premium for someone who retires after in order to make the benefit equitable. C1.3. Evaluation of a Points scheme 65. A “Points� scheme such as that which is in place in Germany functions as follows: 34 In Germany, a year’s contribution at the average earnings of contributors earns one pension point. Contributions based on lower or higher earnings earn proportionately less or more pension points. At retirement, the pension points of every year are summed up and the sum of pension points is multiplied by a “pension-point value.� The pension point value is valid for newly retired and already retired pensioners and is adjusted annually in line with gross wages and can reflect additional factors including changes in the system dependency ratio. 66. In Mongolia, a “Points� scheme with the same parameters as the defined-benefit scheme evaluated above would result in identical system finances and expected benefits. For example, if the target replacement rate for a full career worker contributing for 40 years is set at 40% of the average wage, the system finances and expected benefit outcomes would be identical to the reformed defined-benefit scenario (See Figure 20, Figure 21, Appendix 5, Figure 38 and Figure 39). 67. The differences between a “Points� scheme, a defined-benefit scheme and an NDC scheme are therefore subtle and limited, provided that the defined-benefit scheme has lifetime wage base and wage valorization. All three schemes can establish adjustment mechanisms for rebalancing system finances in the face of unanticipated demographic changes and can be structured to create actuarially fair incentives for working longer. Finally, all three schemes face similar tradeoffs in respect of indexation during the payout phase. C1.4. Evaluation of a Funded Defined-Contribution (FDC) scheme 35 68. Replacing the existing NDC scheme for post-1960 cohorts with a Funded Defined Contribution (FDC) scheme was evaluated. 36 Such an FDC scheme was assumed to have the following characteristics: (i) contributions after enactment of the reform would go into a separately constituted pension fund with funded individual accounts and such funds would be actively managed by private investment managers either in Mongolia or abroad; (ii) recognition bonds payable at retirement would be issued for all existing obligations to post-1960 cohorts for years of credited service prior to enactment of the reform; (iii) the 34 See OECD, Pensions at a Glance, 2009: Retirement Systems in OECD Countries, Online Country Profiles, Germany, www.oecd.org/els/social/pensions/PAG . 35 Further discussion of broader issues of pre-funding pension obligations is discussed in Section V. 36 A so-called hybrid scheme has not been evaluated which could be a combination of retaining some of the existing NDC scheme and replacing part of it with an FDC scheme. -34- State Subsidy would cover the so-called transition costs which represent the amount of funds which would be collected from contributions and placed in individual accounts until such individuals retire; and (iv) the annuity in the payout phase would be structured similarly as under the NDC scheme. Additional parameters were the same as under the reformed NDC scheme and other reform scenarios. 37 69. An FDC scheme can provide replacement rates superior to an NDC scheme if the lifetime rates of return on funded balances (after administrative costs and fees) exceed wage growth. Such returns however come with investment risks that can reduce benefits and with potentially substantial administrative costs for account and asset management. Multiple scenarios with different rates of return on pension assets relative to wage growth were simulated. 70. Expected benefits under an FDC scheme are very sensitive to the rate of return the pension fund will earn on its assets (Appendix 5, Figure 40). Assuming a rate of return equal to wage growth, the projected replacement rates under the FDC scheme are similar to those projected in the NDC scheme (with similar increases in the retirement age and the contribution rates - Scenario 2E). If financial markets perform well and assets are invested wisely resulting in a rate of return 2% above wage growth, system members can get up to 70% replacement rates and higher. But if financial markets are not sufficiently developed and assets yield returns 2% below wage growth, benefits can become inadequate with replacement rates well below 30%. 71. With contributions of post-1960 cohorts invested in individual accounts in an FDC scheme, the pension system will have to continue paying benefits to existing pensioners and pre-1960 cohorts once they retire as well as partially finance disability and survivorship benefits. 38 This yields a sizeable increase in the fiscal subsidy required, with projected deficits in the remaining PAYG scheme practically doubling to 5% of GDP as a result of the transition costs (Appendix 5, Figure 41). 37 No minimum pension guarantee is provided for post-1960 cohorts, the retirement age is gradually increased to age 65 for men and women and the contribution rate is gradually increased to 19%. Contributions made by the post- 1960 cohort are directed to individual accounts in a separate pension fund. 38 In this scenario, disability and survivorship benefits are assumed to remain at pre-reform levels and paid out of the PAYG component. Individual account balances in the funded DC component of those who become disabled or die are turned over to the PAYG component which tops up benefits to the levels guaranteed by the law if the balances are not enough to finance them or keep the difference if they exceed the amount needed. -35- Box 1: Design Issues for Funded Defined Contribution Schemes FDC schemes have a number of important design dimensions: • Are all of funds contributed and accumulated in a Pension or Social Security Fund set aside to back up the future benefit entitlements of members? Generally in FDC schemes all funds accumulated are the property of the member and therefore are used to fully support a pension benefit. • Who bears the risk and receives the return on pension fund assets during the accumulation phase? Generally individual contributors bear most or all of the risk in funded defined contribution schemes and receive most or all of the returns (net of management and administrative costs). • Who bears the responsibility for the selection of a pension fund manager? Selection of fund managers may reside with the pension managing agency, employers, employees or other third parties. • Who bears the risk and receives the return on pension fund assets during the payout phase? There are several different design arrangements for the payout phase so that investment risks, longevity risks, inflation risks and liquidity risks may be the responsibility of the sponsor, management agency, employer, employee or some combination thereof. If an annuity contract is purchased by the beneficiary, then the insurance company selling the contract bears most of the risks. 72. The risks of establishing an FDC scheme appear to outweigh the benefits in Mongolia. The additional returns in excess of wage growth as provided for in the current NDC scheme would be modest when compared to the more substantial investment risks of such a scheme and downside risks would be borne by contributors’ replacement rates which could impact credibility and compliance. Further, sufficient depth, regulatory infrastructure, contestability and other enabling conditions do not exist in the domestic financial markets which can support such an arrangement and will likely not exist for some time to come. Finally, utilization of international markets for such reserve accumulations still presents substantial principal-agent challenges and risks for which Mongolia does not have a legal, regulatory and supervisory structure to properly manage. C2. Pension Insurance for Permanent Disability 73. Mandatory disability insurance aims to provide a source of income for those who lose productive capacity during their work lives and face a significant loss of earnings. Disability insurance has many of the same incentive problems as other forms of insurance. The rationale for a mandatory scheme is that individuals underestimate the probability of such disability, and often those that will try to insure will be those with the greatest probability of becoming disabled (adverse selection). A moral hazard problem is that the benefit can lead individuals to more risky behavior. Moreover, many developing countries have very limited markets for private voluntary disability insurance. Mandatory disability insurance is therefore a means of addressing these incentive and market failure problems. 39 The benefits can also reduce labor participation as those who can qualify for 39 See Estelle James, Disability Social Insurance, draft mimeo, 2010. -36- benefits have less incentive to work. Finally, the benefit costs can be high, particularly in light of the challenges of objectively assessing eligibility certification. 74. The existing pension insurance scheme has two forms of disability benefits, one for total invalidity (based on 70% and more loss of working capacity) and one for partial invalidity (from 50 to70% loss of working capacity). Benefits are: (i) a replacement rate of 60% for total invalidity; and (ii) a replacement rate of 60% times the percentage of loss of capacity for work, for partial invalidity. A 20 year vesting requirement has to be met for full benefits for pre-1960 cohorts and 15 years for post-1960 cohorts. Minimum benefits are the same as for old age pensions. 40 Eligibility certification is determined by the Medical Labor Accredited Commission, and the decision on whether cases are relevant to employment injury is assigned to the Employer’s Investigating & Registering Commission. Benefits are provided from when eligibility is certified until rehabilitation. 75. Contributions for disability benefits are not explicitly earmarked. From 1999 to 2008, 4 percent of wages were allocated for disability and survivorship insurance as well as for administrative costs. 41 For 2011, an estimated 15% of total old age, disability and survivorship pension insurance expenditures went to disability benefits. The 53,400 disability beneficiaries represented about 18% of total beneficiaries for old age, disability and survivorship benefits. Benefits which went to pay disability benefits in 2011 represented about 22% of total contributions. If disability was to be a separately financed benefit, it would therefore require a dedicated contribution rate of 3.1% of wages. Average benefit levels of the disabled represented about 29.4% of the average covered wage. 76. Although disability benefit design needs further evaluation and analysis, some design principles should be considered: • The benefit should be structured as a pooled insurance product from an earmarked contribution and the design and cost of the benefit should be aligned with both initial and periodic actuarial valuations based on changing demographics and benefit incidence. • The benefit formula could include both a flat minimum pure insurance product which is not linked to wages and an income-sensitive component. This would strengthen incentives for accurate wage reporting. 40 For pre-1960 cohorts, the minimum is 75% of the minimum wage and for post-1960 cohorts, 20 percent of the national average wage, plus an additional 0.5 percent of the average wage for each additional service year beyond the minimum of 15 years. 41 The 2008 amendments do not specify what percent of resources should be earmarked for disability, survivorship and administrative costs. Evidently 4% of the total contributions or 0.56% of wages have been earmarked for these costs though the amount allocated to NDC account balances was less than 13.44%. -37- • The minimum benefit should be sufficiently low to discourage potential beneficiaries from applying as a result of other non-disability sources of adverse circumstances while at the same time aiming to provide a minimum level of adequacy. • Upon reaching retirement age, the beneficiary’s disability benefit should be replaced by a retirement benefit. • The benefit should be annuitized and subject to automatic price indexation. • Vesting should be of sufficient length to discourage moral hazards. 77. The choice of the disability benefit level involves tradeoffs between adequacy and affordability. As a starting point for considering different contribution and benefit levels, different benefit levels were simulated into the future. 42 The results suggest that a 2% contribution rate can achieve a benefit of about 19% of average wages for a man though this declines to about 10.0% of average wages in the long term. 43 A 19% replacement rate would represent less than half of the anticipated old age replacement rate provided by the Tier II pension insurance scheme. This is merely a starting point for consideration of disability benefit design and level and does not anticipate the costs of a minimum benefit, different design features and important incentive effects of design options considered. 78. There are also certification, recertification and related design and implementation issues which impact benefit incidence and costs: • The process of eligibility certification, the composition of the medical review committee and the requirements and procedures for recertification are all essential to the incentives to receive disability benefits. 44 • One way in which countries have limited costs and motivated individuals back into the workforce is by incentives for employers to place disabled workers back into other positions. 45 • Disability and work-injury prevention is also another area which is essential to limiting incidence and cost. 42 Past incidence of disability and costs (as a proportion of contributors) were assumed to continue into the future. Once the authorities have agreed upon a target benefit level options can then be simulated which include both the minimum and the income sensitive benefits. 43 This assumes a continuation of the same disability incidence per contributor observed in 2008 projected into the future. The total wage bill is projected to grow in line with the baseline and reform PROST projections. 44 This report cannot judge the impartiality of the Medical Labor Accredited Commission. However, some countries establish have eligibility certification and recertification done by a committee that includes representatives that can have an incentive for cost containment. In countries with private provision of mandatory disability insurance, the eligibility certification is often done by a committee that includes both worker and insurance company representatives which balances groups with different interests. In Mongolia the insurance sector is not well developed so such an approach may not be applicable. 45 See Estelle James, Disability Social Insurance, draft mimeo, 2010. -38- C3. Pension Insurance for Survivorship 79. Survivorship pension insurance programs are designed to support the families when a wage-earner has died and the spouse and other dependents face a substantial loss of income. Survivor benefit programs basically have two goals: preventing poverty and smoothing consumption levels over the lifetime of surviving dependents. Survivor benefit programs are therefore often designed to support surviving dependents unable to work productively. The rationale behind mandatory survivorship programs is that myopia and inability to understand risk faced and/or solutions available may lead workers to underinsure and under-save for their families. Mandatory programs for survivors supplement or supplant insufficient voluntary savings and insurance, helping to achieve an optimal lifetime consumption pattern for the family. 80. Survivors of contributors to the Pension Insurance Fund born after 1960 receive a benefit of 40% of the monthly average wage in the last three years for one dependent increased by 10% per each extra member of the household. The pension should not exceed 60% of the monthly average wage. A spousal survivor would receive such a benefit from the date of passing of a spouse up to his or her retirement age. Young children receive survivorship benefits up to age 16 or if students age 19, or if incapacitated for the period of loss of working capacities. 46 The minimum benefit of survivors under the NDC scheme is the same as for old age pensions. 81. As with disability insurance, contributions for survivorship benefits are not explicitly earmarked. In 2011, there were an estimated 21,800 survivorship beneficiaries (including an estimated 1,700 children dependents) and expenditures were about 7.3% of the total pension insurance expenditures which translates into a pay-as-you-go contribution rate of about 1.5% of covered wages. 47 82. Just as with disability, the choice of survivorship benefit levels involves a tradeoff between adequacy and affordability. As a starting point for considering different contribution and benefit levels, benefit levels were simulated into the future. 48 The results suggest that a 1.5% contribution can achieve an average benefit of about 35% of the 46 Beneficiaries can include grandchildren, brothers and sisters under 16 without caregivers; grandchildren, brothers and sisters disabled or got disabled prior attaining age 16; parents over retirement age or disabled parents, spouse or grandparents, brothers and sisters without caregivers; any of parents or spouse not working and caring for children under 8, or grandchildren and younger brothers and sisters, also family dependents of the deceased who was on receipt of retirement or invalidity pension and who totally lost capacity for work in months preceding the death; step-parents; step-children not receiving alimony from their own parents; family dependents of the disappeared. 47 As there are multiple survivors per deceased worker, one cannot attribute how many contributors this figure applies. Moreover SIGO does not have historical data on death rates of contributors. 48 Current mortality incidence (as a proportion of each cohort) was assumed to continue into the future. Once the authorities have agreed upon a target benefit level one can then simulate options which include both the minimum and the income sensitive benefits. -39- average insured wage for all survivors. 49 This is merely a starting point for consideration of survivorship benefit design and level and does not anticipate the costs of a minimum benefit and different design features. It does suggest however, that a 1.5% minimum mandatory contribution may be sufficient for a modest annuitized benefit for direct dependent survivors. In considering the reform design, it is important to bear in mind that the size and duration of the survivorship benefit will impact the incentives for workers to seek other forms of life insurance or savings and of course will impact the incentives of survivors to seek alternative sources of income. Moreover, the higher the contribution rate for the pooled flat benefit, the greater will be the incentive to underreport wages. 83. Although survivorship benefit design needs further evaluation and analysis, some design principles should be observed: • The benefit should be structured as a pooled insurance product from an earmarked contribution. • The benefit formula could include both a flat minimum which is not linked to wages and an income-sensitive component. This will strengthen incentives for accurate wage reporting. • The benefit should be annuitized and subject to automatic price indexation. • Periodic actuarial projections should ensure that the long-term benefits are aligned to long-term revenues and reflect changing demographics and mortality experience. D. Tier III – Supplementary Voluntary Pensions Savings 84. Occupational and personal savings arrangements are an essential part of providing old age income protection. The rationale behind these instruments includes: (i) the dramatic aging of the population and increases in old age system dependency rates suggests that the pension insurance system will likely only be able to provide retirees with replacement rates of 35-45% of pre-retirement income in the long-term therefore necessitating a supplemental source of income support for many retirees; (ii) changes in the household structure and reductions in co-residence can result in an increasing number of people that need alternative savings opportunities for old age income protection; (iii) voluntary occupational schemes can provide a framework to compensate for the uneven level of benefits between different types organizations and enterprises; (iv) occupational schemes can enhance the economic efficiency of enterprises and Government employers because the structure of the benefits can be tailored to increase the willingness of employers to invest in training and skills development; (v) voluntary savings arrangements may be better suited to the uneven income streams of many herders and members of the 49 This assumes a projection of mortality experience according the UN population projections mortality tables. The total wage bill is projected to grow in line with the baseline and reform PROST projections. -40- informal sector; and (vi) privately managed pension funds can support the development and stability of financial markets which in turn can be supportive of overall economic development. 85. In spite of their importance, occupational and personal savings arrangements are largely non-existent in Mongolia. Only companies which are registered with the tax authority as authorized pension annuity providers can receive a deductibility of contributions for tax purposes and this process of authorization evidently has yet to be facilitated. Moreover, the provision of annuity products has only just begun amongst life insurance companies. Finally, some companies and organizations have offered their own supplementary pension plans though such plans have not been regulated and not subject to preferential tax treatment. The absence of a regulatory framework and favorable tax treatment understandably has discouraged the development of such occupational and personal savings arrangements. 86. The proposed design features for occupational and personal savings arrange- ments are summarized in Figure 22. Some of the proposed features are: (i) The schemes should be defined contribution and fully funded to provide a framework for portability of benefits as well as a transparent means of verifying pension entitlements through account accumulations; (ii) account accumulations should be fully portable; (iii) account and investment management should be through licensed and regulated financial intermediaries independent of and on an arms-length with the sponsoring employer; (iv) account balances above a level to be determined should in principle be provided in an annuitized form so that individuals can have a means of managing longevity risks; (v) a strengthened regulatory framework and supervisory capacity is essential to ensure employer compliance with minimum standards and adequate consumer protection; and (vi) a uniform framework for record-keeping agency is critical to ensure accuracy and transparency of records and reduce administration costs. 50 50 Some countries have employed the use of a central record-keeping agency. -41- Figure 22. Proposed Occupational and Personal Pension Arrangements Issue Provision/Parameters Architecture • Defined contribution. Funding • Fully funded. Tax treatment • Consistent tax treatment (e.g. Exempt, Exempt, Taxable or Taxable, Exempt, Exempt). • Employee contributions provided personal tax deductibility only up to a limit (both as a percent of household taxable income and a cap on the total permissible deduction). • Employer contributions should also be subject to limits for deductibility. Financing • Through individual, employer and/or employee contributions. Withdrawal and • Withdrawal provisions (lump-sum, phased-withdrawal and annuities) to be annuitization developed. provisions • Withdrawal should be permitted at a retirement age close to that of social insurance though earlier in the case of unemployment or other demonstrated needs. • Age of withdrawal should be linked to individual tax treatment Regulation • Strengthened regulatory framework for investment management, custodianship, and account management. • Revision of role, responsibilities and accountabilities of a Trustee; provision of governing standards. • Regulatory framework and institutional infrastructure supporting a private annuity market to be developed. Supervision • Strengthened supervisory capacity. E. Transition Arrangements 87. This section elaborates transition options for smoothing replacement rates between pre- and post-1960 cohorts and examines how to replace current minimum pension guarantees with the proposed Targeted Social Pension. 88. One transition option for consideration is to establish a 10 year program to smooth the abrupt drop in replacement rates for post-1960 cohorts as follows: (i) For women retiring between 2015 and 2024 and men retiring between 2020 and 2029, it is possible to calculate benefits both according to the applicable NDC benefit formula and according to the formula applicable for pre-1960 cohorts. A graduated portion of difference between such benefit calculations could be provided through a benefit supplement according to a schedule as suggested in Table 7 below; (ii) such a benefit supplement could be provided as a tax-financed benefit supplement separate from and in addition to benefits provided under the Pension Insurance scheme. In this way, the supplementary benefit would not become an entitlement for future beneficiaries; and (iii) the benefit should be indexed in the same way as other pension insurance benefits. Such a supplementary transition benefit has the appeal of avoiding the severity of a decline in replacement rates for post-1960 cohorts during the transition period. At the same time, enactment of a targeted social pension would protect the bottom half of the income distribution of retirees from elimination of the minimum pension guarantee. In this way, a tax-financed benefit supplement would be provided at the same time in which parametric reforms gradually increase replacement rates (see below). -42- Table 7: Possible Benefit Enhancement to Smooth the Transition for Pre- to Post-1960 Cohorts Year 51 Supplementary Benefit – Proportion of the difference between NDC and DB benefits 1 90% 2 80% 3 70% 4 60% 5 50% 6 40% 7 30% 8 20% 9 10% 10 None 89. Another transition to consider is how to replace the minimum pension guarantee with the proposed targeted social pension. Replacing the minimum pension guarantee for post-1960 cohorts with the proposed targeted social pension could materially affect the replacement rate for retirees with low incomes or limited years of contributions or both. The impact and the need for transition arrangements would depend upon the relationship between the benefit provided by the minimum social insurance pension guarantee on the one hand, and the maximum benefit under the targeted social pension on the other. 52 If the level of the benefit chosen roughly conforms to the average pension guarantee for post-1960 cohorts, no transition mechanism may be needed. If however, there is a divergence in the parameters between the minimum pension guarantee on the one hand and the targeted social pension, on the other, then a transition program would have to be developed. Parameters for such a transition could be similar to that proposed for a supplementary benefit to smooth the transition for post-1960 cohorts above. In this way, over a 10 year period, any differences between the benefit provided under the targeted social pension and the pension insurance scheme minimum could be gradually smoothed. 51 The years would start consecutive from 2015 for women and 2020 for men. 52 At end-2010, 60% of the Minimum Living Standard of MNT101,400/month would be MNT60,960. By comparison, the minimum pension guarantee for men born after 1960 would be 20% of the average wage (in 2020) plus 0.5% per year for each year of service after 15 years. At 2010, this benefit would be about 22% * MNT310,000 or about MNT68,200/month, an amount almost identical to 60% of the MLS. -43- IV. Addressing the Special Needs of Herders and the Self-employed A. The Coverage Gap 90. Mongolia lacks an effective retirement income strategy for those who are currently uncovered by contributory pension insurance including herders, the self- employed and non-wage workers. Although the overwhelming majority of old age retirees currently receive benefits from the Pension Insurance scheme, only about half of the economically active population was covered in 2009 (Table 8). 53 Herders, who are estimated to represent about a third of the labor force, and the self employed, who represent about 15 percent of the labor force, are both exempted from making mandatory contributions to the scheme. Table 8: Social Insurance Contributors, Beneficiaries, Economically Active Population Individuals 1995 2000 2005 2006 2007 2008 2009 Number of insured people covered by pension, unempl. & ins. against employment injury and occupational diseases 409,100 381,400 367,936 436,385 484,848 547,161 543,884 Compulsory insured 395,400 363,900 338,930 410,088 433,710 484,608 492,163 Voluntary insured 13,700 17,500 29,006 26,297 51,138 62,553 51,721 Number of insured people covered by health insurance 1,992,400 2,030,200 1,959,400 1,905,742 2,101,734 2,233,657 2,122,939 Total of beneficiaries 284,277 254,394 264,177 267,520 273,233 279,927 279,306 Pension beneficiaries 222,816 179,668 179,959 183,541 188,500 193,378 196,789 Population, labor force Population /in thousands/ 2,251,300 2,407,500 2,562,400 2,594,800 2,635,200 2,683,400 2,735,800 Economic active population /in thousands/ 812,700 847,600 1,001,200 1,042,800 1,054,000 1,071,513 1,044,400 Employed 767,600 809,000 968,300 1,009,900 1,024,100 1,041,700 1,006,300 Coverage of Economically Active Population 50.3% 45.0% 36.7% 41.8% 46.0% 51.1% 52.1% Pension System Dependency Rate 54.5% 47.1% 48.9% 42.1% 38.9% 35.3% 36.2% Source: SIGO Statistical Yearbook of Social Insurance, 2010. 91. The coverage gaps affecting herders, the self-employed and non-wage workers emerged as a by-product of the economic reforms of the early 1990s but only more recently are becoming a significant problem. Prior to these reforms, most economic activity was conducted through cooperatives, collectives and state enterprises, where participants earned pension credits in the course of their work. For example, nomadic herders were organized into cooperatives, the cooperatives owned the livestock, and the herders were paid (and awarded pension credits) to care for the animals. After the reforms, herders, farmers and many urban entrepreneurs became independent economic actors who were no longer paid a salary and no longer automatically earned pension credits through their work. Those with at least 20 years of credits under the Pension Insurance scheme qualify for a full pension, even if it is only at the minimum pension level. 53 There are a number of concerns over the quality of this administrative data though no alternative sources of data from which to validate it. -44- Men turning 60 in 2011 were in their early 40s when the economic reforms were introduced and are likely to have had 20 years of service prior to the implementation of the reforms. Those turning 60 beginning in 2015 or so as well as women turning 55 beginning about now will likely not have had enough time prior to the reforms to earn a full pension benefit. For a few years, newly retiring workers may be eligible for partial minimum pension by virtue of having 10 years of credit under the old system. Even that benefit will no longer be available to most women reaching the retirement age after 2014 and most men reaching retirement age after 2019 where 15 years of contributions are required for pension eligibility. 92. Herders and other self-employed have the option of participating in the Pension Insurance scheme by making voluntary contributions equal to 10% of the prevailing minimum wage. At end-2010, that implied a contribution of some MNT 10,800 per month or just under MNT 130,000 per year. Only about 5% of the economically active population makes such voluntary contributions (about 52,000 or 5% of the economically active population in 2009). The age composition and level of service of voluntary contributors suggests that many are participating with the explicit intention to meet the 20 year vesting requirement for a defined-benefit pension for pre-1960 cohorts or the 15 year vesting for post-1960 cohorts. Those who reach retirement age without being eligible for a full or partial pension insurance benefit and may qualify for a Social Welfare Pension. B. Policy Options 93. Three options are commonly considered for closing the coverage gap to improve the retirement incomes of herders, the self-employed and other non-wage workers: a) extending coverage on a mandatory basis, either to the pension insurance scheme or to a new scheme designed specifically for herders and/or other informal sector workers; b) encouraging greater participation in a voluntary retirement income scheme, by creating an entirely new matching defined-contribution (MDC) scheme; or c) integrating the existing social welfare pension into an improved targeted social pension that provides more adequate benefits to the elderly with very limited pension incomes from other sources. B1. Mandating contributions for herders and other informal sector workers 94. Mandating participation for herders and other informal workers suffers from two major shortcomings: the extremely low incomes among much of the target population and the difficulty of enforcing any mandate. Most herders and informal sector workers have very limited incomes and/or assets and the incomes they do have are subject to substantial volatility. The government estimates that about two-thirds of the herders own -45- fewer than 100 animals, a number it considers to be too small to provide a reasonable living. Under its State Policy on Herders (adopted in June 2009), those with fewer than 100 animals are to be encouraged to find another line of work. 95. If participants in the current voluntary scheme were required to contribute to a mandatory program at the current combined employer-employee rate of 14%, their annual contribution would rise from MNT130,000 per year to MNT180,000. Many of the owners of fewer than 100 animals and most of those now in the urban informal sector are unlikely to have the cash resources needed to make these kinds of pension contributions. 96. In addition, any mandate of pension coverage would be difficult to enforce. These workers are already required to participate in the health insurance program by paying annual premiums of MNT 8,040. The annual health insurance premium is about one-twentieth of the minimum voluntary pension contribution. The government estimates, however, that only about 94,000 of some 380,000 herders pay the health insurance premium, even at such a low premium level. 97. Collecting a much larger pension contribution would obviously pose a significant challenge. Herders operate largely outside of the cash economy, marketing animals and animal products a couple of times a year, often through informal marketing channels. The government has little information about the cash incomes of individual herders and, as illustrated by the health insurance experience, has few tools to enforce a participation requirement. 98. In the past, alternative schemes have been developed that would scale pension contributions to the number of animals owned by each herder. That approach might be somewhat easier to enforce than an attempt to scale contributions to income, since the size of each herd is more easily observed than is the herder’s income (though disclosure of herd size could also be manipulated). No scheme can overcome the barrier presented by the very limited resources of the majority of herders, however. Under such a scheme, producing a reasonable pension to smaller herders would require substantial cross subsidies from owners of large herds plus a generous subsidy from the government. Moreover, the idea of scaling contributions to herd size runs counter to recent trends, at least in local government finance. In recent years, the use of such “hoof taxes� has been abandoned by most local governments. 99. While a mandatory program does not seem to be a promising approach for the majority of herders at the present time, it might in the future prove a practical way to improve the retirement incomes of at least the largest herders, self-employed and non- wage workers. As noted, the official government policy is to encourage the smallest -46- herders to find other means of support, including becoming employees of the larger herders. The government also intends to encourage those with medium sized herds to associate with cooperatives that can facilitate purchasing, marketing and management of both the animals and the pastures. It also wants to develop more formal marketing channels capable of increasing the quality of the products produced. 100. If these initiatives prove successful, they could make it easier at some future date for the government to enforce mandatory participation in a Pension Insurance scheme, particularly if the mandate is restricted to the larger herders, self-employed and non-wage workers. The government might require participation, but exclude any herder whose income was below a specified amount, perhaps the minimum wage or some stated multiple of that wage. This would restrict the mandate to those with the resources needed to comply. It also would focus on a part of population more likely to use the new more formal marketing channels, which would help the government to monitor compliance. It may also be possible at some future time to extend mandatory coverage to those urban self-employed who also have higher incomes and more linkages to the formal sector. 101. Restricting mandatory coverage to those with the largest herds and self- employed with highest incomes would allow the government to require contributions at the same rate as other contributors. Focusing only on those with higher incomes would also make it more likely that the resulting pensions would be larger than the minimum. Taken together, these two factors would greatly reduce the amount of any subsidy that the government would need to make to assure decent retirement incomes for these particular workers. Government subsidies, either in the form of a new universal program or one of the other options considered here, could then be directed at the less successful self-employed and those herders with relatively modest resources. B2. Expanding the Voluntary Pension Program and Introducing Matching Contributions 102. A second approach to closing the coverage gap would be to expand or restructure the voluntary pension program, but this is likely to face the same barrier of limited resources among much of the target population. The current program of voluntary pension contributions already involves a substantial subsidy. Contributions of 10% of the minimum wage over a twenty year period will generate a lifetime payment (from age 55 for women and age 60 for men) equal to the minimum pension. Since the minimum pension is currently about 75% of the minimum wage, the program essentially provides a 75% replacement rate after 20 years of contributions at a lower contribution rate of 10%. Under plausible assumptions, the participants’ actual contributions currently amount to -47- less than half – and as little as 15 percent – of the total pension benefits they can expect to receive. 54 103. Under the Social Welfare Law, those aged individuals currently not entitled to a benefit from the Pension Insurance scheme are eligible for a social welfare pension, which at end-2010 was set at MNT 53,800 per month. Since the minimum benefit under the Pension Insurance scheme at end-2010 was MNT 105,000 and receipt of the national pension disqualifies one from the social welfare pension, participation in the voluntary contribution program can be said to offer only an additional MNT 51,200 (the difference between the two amounts), rather than the full MNT 105,000. Viewed this way, the fraction of the benefit financed by participant’s contributions doubles and the amount financed through subsidies are reduced proportionately. 55 104. Despite the sizable subsidy, relatively few people participate voluntarily. Those who make voluntary contributions tend to be people nearing retirement age who already had most of the 20 years of service that they need for vesting for full entitlement. 56 In contrast, few people in their 30s and 40s are trying to generate pensions entirely from voluntary contributions. In other words, the program is being used to close modest coverage gaps rather than to generate entirely new entitlements. Current voluntary participation is less than 10% of the contributors to the pension system while herders, the self-employed and informal sector accounts for almost 50% of the work force. 105. The current arrangements for voluntary pension contributions include several features that are likely to discourage participation. One is the interaction with the social welfare pension, which offsets a good part of the subsidy otherwise given to voluntary contributors. Another is the relatively inflexible contribution arrangement. A third is the locking in of the contributions until retirement age is reached. And a fourth is that the size of the subsidy is hidden and probably not fully appreciated by those eligible to participate. Addressing each would make participation more attractive and might increase the participation rate, particularly among those in their 30s and 40s. 54 Under these assumptions, if the full minimum pension rises in the future at the same rate as wages rise, twenty years of contributions (plus interest earnings) would amount to only about 19 percent of expected benefit payments. If it rose in the future only at the rate that prices increase, contributions plus interest would account for 44 percent of expected payments. The key assumptions are that real wage increases average 4.6 percent per year, that balances earn interest at the rate of 5.1 percent per year and that retirees draw benefits for an average of 17.2 years. For those who contributed for only 10 years and receive the lower minimum pension, contributions would amount to between 14 and 21 percent of expected payments at the lower minimum pension level. Contributors also qualify for disability and survivor benefits, which further enhance the value of the package and therefore the size of the implicit subsidy. 55 To be exact, the fraction of the benefit financed by contributions increases from 19 percent to 39 percent, assuming wage indexing or from 44 percent to 86 percent, assuming price indexing. 56 This is based on interviews with officials in Tov aimag. -48- 106. The structure of the current voluntary pension also discourages participation in other ways. The current structure makes no accommodation for irregular incomes, a common characteristic among informal workers and those that sell agricultural products. A more attractive structure might be to introduce greater flexibility by allowing smaller contributions in unfavorable years to be offset by larger contributions in more favorable years. In addition, no provision is made for dealing with fiscal emergencies, a common problem among agricultural and urban informal workers who have few reserves or other forms of insurance. An option that would address this problem would be to offer voluntary contributors the ability to borrow back a portion of their past contributions in cases of financial emergency. 107. The size of the State subsidy could be made more transparent by substituting the current voluntary arrangement with a matching defined contribution scheme which would make the subsidy explicit. If the calculations here are accurate, the current level of subsidy is equivalent to about a two for one match – a government payment of MNT 2,000 for each MNT 1,000 deposited by the worker. A more transparent subsidy may well encourage participation though it still likely only to capture the higher income self- employed. 108. Substituting voluntary contributions to defined contribution accounts instead of the current voluntary contributions to a defined-benefit scheme would address two of the other issues noted. The defined contribution arrangement can be much more flexible about variations from year to year in the amount contributed to the plan. Such flexibility is a bit of a challenge under the defined benefit plan because an exact count of the years of service is an important determinate of benefit eligibility. Such a count is not important under a defined contribution plan, since the benefit at retirement simply reflects the balance in the account. The defined contribution approach also makes it somewhat easier to allow emergency loans against the balance in the account. An outstanding loan balance can simply be deducted from the account balance at retirement. 109. Substituting a defined contribution approach and incorporating these kinds of changes could make the voluntary program more attractive and increase participation. International experience suggests however that the preferences and economic circumstances of herders, the self-employed and informal sector is still likely to substantially limit the growth in coverage that can be realized from the reforms proposed. Most of the participants (and, therefore, most of the recipients of the subsidy) would undoubtedly be the proportion of the target population having the most resources. If the coverage were to increase almost three-fold from 52,000 to perhaps 200,000, this would still leave over 60% of the target population uncovered. Such an approach will therefore likely have a meaningful impact but will need to still be supplemented by additional -49- measures to reach the poorest elderly who will likely still not have sufficient sources of income support in old age. B3. Integrating the Social Welfare Pension into a Targeted Social Pension 110. A third approach is to replace the Social Welfare Pension with a Targeted Social Pension. This is of particular importance for herders, the self-employed and non-wage workers, many of which may not establish sufficient retirement savings to protect against poverty in old age. As suggested, a key consideration is providing a Targeted Social Pension which is both adequate to protect against elderly poverty while at the same time modest enough to still ensure strong incentives for individuals to contribute to their own retirement. 111. A major attractiveness of the Targeted Social Pension is that it can be structured to provide basic income protection for the poorest elderly and designed to complement benefits from matching defined-contribution and NDC accounts. In this way, elderly coverage gaps are much more systematically addressed and a more consistent framework is applied to the elderly across the income spectrum. This is the approach suggested in the design for the Tier I Targeted Social Pension in Section III.B above. B4. A recommended approach 112. The recommended approach draws from each of the three approaches above. It is suggested to: (i) Establish regulated individual retirement savings arrangements in accordance with the development of a Tier III supplementary voluntary pensions savings; (ii) Elaborate a plan for a matching defined contribution scheme for post-1960 cohorts not participating in the Pension Insurance scheme; (iii) Elaborate a plan for introduction of mandatory contributions to the largest herders, self-employed and non-wage workers; and (iv) Introduce a targeted social pension with the characteristics and features earlier elaborated. B4.1 Voluntary Savings Arrangements 113. Regulated voluntary pensions savings arrangements are essential for many of the needs of herders, the self-employed and non-wage workers. Since it takes many years to build up sufficient savings to support a meaningful retirement benefits, individuals must have sufficient trust in the institutions managing the resources and in the regulatory and supervisory authorities to set aside resources for many years of their lives. Long-term -50- savings for retirement involves considerable risks to the contributor. In this way, building up a regulatory framework, supervisory institutions and financial and data management institutions that embody the public trust are essential building blocks for individuals with volatile incomes to set aside scarce resources for future retirement. 114. The volatile nature of incomes and liquidity needs by herders, the self-employed and non-wage workers also suggest that some will utilize well-regulated voluntary savings arrangements. Safety and security of retirement savings is of great importance for this population. B4.2 A Matching Defined Contribution Scheme 115. A matching defined contribution scheme which targets the uncovered economically active population is another important instrument to address coverage gaps. This suggested approach is motivated by several factors: (i) The nature of work among herders, farmers, the self-employed, and informal sector and variability of employment relationships makes it difficult to mandate participation for these types of workers (as indicated above); (ii) the uncertainties and volatility of such non-wage incomes suggests that an incentive to lock up their savings in a pension scheme during their working lives is likely necessary; (iii) a notional individual account accumulation provides the possibility of crafting rules for selective borrowing of some of the accumulation during people’s working lives in the face of specified shocks; and (iv) the design has both a simplicity and compatibility with the existing notional individual account arrangements under the Pension Insurance scheme to facilitate implementation, understanding among participants, fiscal planning, and allocation. 116. The objectives of such a scheme would be: (i) to provide framework for pension savings for non-wage rural and urban residents, including herders, farmers, the self- employed, temporary workers, the informal sector and the unemployed; (ii) to use a state subsidy as a means to creating incentives for individuals to save a minimal level of retirement benefits by matching individual contributions; and (iii) to use the state subsidy as a means of reducing coverage gaps in retirement by increasing pensions savings during individuals’ work lives. A summary of some of the key design features is indicated in Figure 23 below. -51- Figure 23. Possible Parameters of a Matching Defined-Contribution Scheme 1. Participation • Participation should be voluntary for non-wage workers not contributing to the Tier II Pension Insurance scheme 2. Qualifying • Retirement age: same as for the Tier II Pension Insurance Scheme conditions • Immediate vesting – Qualification for benefits at retirement age. 3. Benefit • Benefit formula – The same annuitization benefit formula applicable to the Tier II NDC design scheme would be applicable. • Benefit indexation - Same indexation applied to Tier II NDC annuities would be applicable. 4. Contributions • Minimum worker contribution amount to be determined. The amount needs to be low and subsidies enough to be affordable to the target population while at the same time high enough to provide for a basic subsistence in retirement after a full career of contributions. • Ratio of matching contribution subsidy to worker contribution. A 1:1 match is an often-considered policy choice if only for simplicity. • Multiple options for the frequency of contributions should be offered to (monthly, quarterly, annually). 5. Financing, • Accounts can be notional (as NDC accounts) with the same notional as the NDC accounts subsidies and • State subsidy can be paid at retirement even though earmarked into a notional account. interest rates 6. Management • In principle, it seems that the SIGO can manage MDC accounts in precisely the same way that NDC accounts are managed, with linkages in the accounting and disclosure 117. Determination of the contribution amount which would be “matched� should be calculated based on the assumption of a minimal work history and a minimum target benefit. A useful target is to mobilize sufficient individual savings matched with government contributions to protect individuals against poverty in old age. The minimum and matching contributions can then be determined based on the minimum targeted benefit level. It may also be useful to consider the target benefit as a proportion of per capita income, keeping in mind the tradeoff between affordability of the contributions and adequacy of the benefits. Due consideration is also needed to the incentive for substitution between voluntary contributions subject to matching subsidies on the one hand, and mandatory contributions to the Pension Insurance scheme on the other. 118. A stylistic illustration of monthly benefit levels produced by different worker contribution rates questions of the affordability of a basic benefit under such an MDC scheme (Figure 24). The figure suggests that it would take at least 18 years for a worker to building up sufficient savings for a benefit of 51,800 Tugriks per month (the amount of the Social Welfare Pension at end-2010) assuming a contribution of about MNT 9,900 per month (9.25% of the minimum wage or 3% of the average wage), 4% real growth in wages, and a 1:1 matching contribution subsidy provided by the government. Other scenarios require longer periods of contributions to support such a benefit level. Such simulation suggests that a contribution level which is affordable to most members of the uncovered population can only provide a minimal subsistence in retirement after a full career of contributions. In addition, a contribution of MNT 9,900 per month would be almost 15 -52- times the contributions of MNT 670 required under the national health insurance which raises serious questions of affordability for much of the target population. This simulation therefore suggests that benefits under such an MDC scheme would need to be supplemented by benefits under a Targeted Social Pension in order to ensure basic elderly income protection. Figure 24. Stylistic Examples of Monthly Benefits produced by Matching Contributions (Benefit amounts per month based on different contribution rates and wage growth) 350,000 300,000 250,000 Benefit Amount per Month 200,000 150,000 100,000 50,000 - 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 Years of Contributions Worker Contribution of 2% of average wage, 2% wage growth Worker Contribution of 2% of average wage, 4% Wage Growth Worker contribution of 3% of average wage, 2% Wage Growth Worker contribution of 3% of average wage, 4% Wage Growth Note: The monthly benefit amounts are in present value terms. Worker contributions are flat amounts based on prevailing assumed average wage with an assumed average wage of MNT 330,000 per month in year 1. All scenarios assume a 1:1 match by a state subsidy for the worker contributions. The real interest rate on notional account balances was assumed to be the same as wage growth. The real discount rate used for computing the benefit amount was assumed to be equal to the real wage growth rate. The benefit was structured as a flat annuity based on a life expectancy of 16.7 years which is the projected life expectancy of men at age 60 in 2020. 119. Determining the appropriate rate of subsidized matching contributions should also be based on the anticipated effect such matching contributions will have in mobilizing long-term savings. This will depend on several factors, including available fiscal resources, the elasticity of take-up against different rates of matching, the interaction between the MDC scheme and the Pension Insurance scheme, and the interaction between the MDC scheme and the Targeted Social Pension. There is almost no robust evidence to correlate rates of matching contributions with participation under varying scenarios. For simplicity, a 1:1 match on individual contributions is not uncommon and is proposed as a starting point here. -53- 120. A related important transition consideration would be how to handle those already at or approaching retirement that will have insufficient time as contributors to generate an adequate individual account balance. This is a consideration not only for a matching contribution scheme but for structuring the Targeted Social Pension as well since the adequacy of the Targeted Social Pension benefit should be reduced over time to create an incentive for individuals to voluntarily save in the matching defined contribution scheme. B4.3 Mandatory contributions for the largest herders, self employed and non-wage workers 121. Mandatory contributions to the largest herders, self-employed and non-wage workers are another important measure consider in reducing coverage gaps. This process includes revising policies for mandatory contributions as well as strengthening compliance enforcement. A first policy measure to consider would be to subject the self- employed, non-wage workers with annual incomes above an annual threshold to be determined 57 and herders with herds above a particular size, to make contributions to the MDC scheme, once constituted. A second policy measure to consider would be to develop a second, higher income threshold and to require minimum contributions to the Tier II Pension Insurance scheme for individuals with incomes above such a threshold. 122. Compliance enforcement measures require some care to develop and need to weigh the costs of enforcement against the anticipated improvements in compliance. Some of the compliance measures that countries have adopted to extend coverage have included: (i) Requiring recipients of national and local level government contracts above specified thresholds to register as contributors; (ii) requiring self-employed with municipal business licenses and incomes above a threshold to contribute; and (iii) unifying and cross- checking data between personal income tax collection with the SIGO to identify potentially eligible contributors. B4.4 Introduction of a Targeted Social Pension 123. Introduction of a Targeted Social Pension is aimed at protecting against elderly poverty or vulnerability where mandatory and subsidized voluntary savings and insurance arrangements prove insufficient. Although the three measures suggested seek to mobilize mandatory and voluntary contributions for retirement, the Targeted Social Pension is designed as a safety net of last resort, recognizing that mandatory and voluntary contributions will still prove insufficient to protect against poverty and vulnerability for much of the non-wage population for some time to come. 57 This income level could be the minimum wage. -54- V. Pre-funding Transitions and Future Obligations through a Pension Reserve Fund A. Conceptual Considerations 124. Generally there are three ways of financing pension obligations: (i) payroll tax contributions whereby such funds either are used to pay the benefits of current retirees or are placed into funded individual accounts for contributors to use for their retirement; (ii) other Government revenues can be used to finance pension obligations, such as the State Subsidy currently provided to the Pension Insurance Fund each year in Mongolia; and (iii) assets or funds can be set aside to pay for future pension obligations, often through a Pension Reserve Fund or other type of Sovereign Wealth fund. In the last case, the risk and returns of the Fund is borne by the sponsor, generally the Government, and does not impact the benefit promise to pension insurance contributors. 125. Mongolia’s defined-benefit scheme for pre-1960 cohorts is funded on a pay-as- you-go basis whereby current payroll tax contributions pay the benefits of current retirees. As benefits are higher than contributions, a state subsidy has been required each year. Almost all of the 1960 cohorts will have retired by 2015 (women) or 2020 (men) so that the state subsidy for this scheme will gradually diminish after these dates. It is possible to pre-finance all or part of the state subsidy required for the period 2015-2040 during which such costs would be incurred though the financing vehicle could be retained outside the Pension Insurance scheme through a separately financed fund (See Section B below). 126. The 1999 legislation governing Mongolia’s notional defined contribution scheme for post-1960 cohorts specifies that the benefit formula is based notional account accumulations. Notional accounts receive a notional interest rate equal to three year average growth in covered wages and are not linked to any investment returns on reserves. However, an annex to the legislation suggests targets for pre-funding future liabilities even though the legislation suggests that such funding targets in no way impact the benefits received by retirees. In each year since 1999, the Pension Insurance Fund has disbursed more in benefits than it has received in contributions so that the funding targets specified in the legislation were never realized. B. Pre-Funding Transitions and Future Obligations through a Pension Reserve Fund 127. Several countries have established national pension reserve funds with the intention of pre-financing future pension benefit obligations (See Appendix 3). An essential legal distinction which distinguishes these such funds from a “funded� approach to pension system finances is that the rate Government savings and rate of return on assets -55- in a such a Pension Reserve Fund does not directly impact the pension benefit obligation promised to the worker. In this way, a pension reserve fund is a means of pre-financing future benefit obligations while not subjecting contributors to the investment or funding risks of such funds. 128. One advantage of such an approach when compared to establishing funded individual accounts within the pension insurance scheme is that pension liabilities are not explicitly linked to the returns on such funds. As such, benefits are not reliant upon the capital accumulations of such funds. As a result, if the reserve fund under-performs its desired investment objective or somehow is otherwise diverted for other fiscal purposes, then the retirees are not left with an un-funded or inadequately funded shortfall. 129. The authorities are currently considering its medium-term strategy for the investment of excess mineral royalties. Agreement on a general strategy is important because of the variety of different ways that the mineral royalties might be used. In the absence of a clear plan for their use, Mongolia runs the risk of not achieving the full benefits of a carefully considered long-term strategy. The authorities have indicated that excess revenues will be placed into the existing fiscal stabilization fund until the balance reaches between 5-10% of GDP which is anticipated by about 2015. At such point, the authorities may consider to establish a pension reserve fund and/or one or more commodity stabilization funds as part of a broader strategy. 130. There are several types pension costs that the authorities may want to consider as options to “pre-finance� possibly through constitution of a Pension Reserve Fund: (i) Pre-finance a transition program for post-1960 cohorts to supplement benefits from about 2015 to 2030 during which time the impact of parametric reforms will be able to adjust the average replacement rates of these cohorts back up to a level of about 42%. (ii) Pre-finance the costs of a targeted social pension for a 15-20 year period during which time uncovered workers or workers with insufficient entitlements will have time to build up their notional account balances while many older workers closer to retirement will not have such an option. (iii) Pre-finance the matching contributions for a matching defined-contribution scheme for a discreet period. (iv) Pre-finance the projected fiscal subsidy required for pre-1960 cohorts from 2015 for a period from 2015 to 2040 by which time most expenditures will be eliminated. -56- (v) Pre-finance a portion of pension contributions to offset part of proposed 5% increases from 14-19% in order to much more gradually increase contributions over a 20 year period beginning in 2015. (vi) Pre-finance the anticipated contributions required for disability and survivorship benefits which could be 3.5-6.0% of covered wages for a specified period. (vii) Earmark funds to pre-finance the needs for old-age income security and poverty protection but refrain from specifying the precise intended use of such funds. 131. It is difficult to determine precise costs for each of these potential uses of a Pension Reserve Fund because the reform parameters and transition plans are still under consideration. Range estimates have been provided for those scenarios for which projections have been prepared. 132. Setting up a Pension Reserve Fund involves critical decisions in at least three different areas (See Appendix 3): (i) funding and withdrawal rules; (ii) institutional structure; and (iii) investment policies. Funding and withdrawal rules involve determining the schedule for deposits into the fund and the conditions under which fund assets will be disbursed. Presumably, a tentative schedule for depositing royalty income into the fund will be developed based on consideration of the expected total royalty income available, the competing uses for that income, the foreign exchange consequences of different approaches, and the likely impact of different accumulation schedules on the fund’s ability to offset future pension costs. The accumulation schedule is likely to be a plan that is subject to annual review and adjustment as more is known about actual royalty income and the value of each alternative use. At least initially, the withdrawal plan may involve a prohibition on any withdrawals until a specified future date or until the fund reaches a certain size. Thereafter, it may involve a specified goal as to the fraction of future costs to be met from the fund. 58 133. A variety of institutional arrangements are used around the world to govern and manage Pension Reserve Funds (See Appendix 4). In some cases, assets are managed by the same agency that also manages the pension plan; in other cases, asset management responsibilities are lodged in a separate institution. A number of pension plans in the developing world either operate as provident funds or have reserves that were 58 The relationship between the size of a pension fund and the level of expenditure support it can supply on an ongoing basis is given by the equation: g = (r-e)*R, where g is the level of spending that can be supported from the fund, as a percent of GDP, r is the average rate of return earned on the fund, e is the average rate of growth in total pension expenditures and R is the ratio of the balance in the fund to GDP. Thus if r is 5 percent, e is 2 percent and the balance in the fund is 50 percent of GDP, then the fund can support expenditures totaling 1.5 percent of GDP on an ongoing basis. -57- accumulated in the plan’s early years when benefit payments were low. As a general rule, national pension assets held in developed countries are also managed by separate agencies. There is no hard evidence as to the superiority of one institutional arrangement over another, but international expert opinion probably favors lodging asset management responsibilities in a separate agency for two main reasons. First, asset management requires a set of skills and operating modes entirely different from the skills and operating modes needed to run an effective pension agency. As there is little synergy between the two kinds of activities, there is little to be gained by housing them in the same institution. Second, it is likely that the asset management function can be better insulated from day to day political interference if is lodged in a separate institution that is designed to be somewhat independent. In some cases, particularly when the fund is just beginning and/or not expected to ever be very large, the asset management function can be lodged at the central bank. Where the size of the fund justifies the expense of creating a separate organization, these are usually created with boards of directors composed of individuals who are appointed to fixed terms by the political leaders, responsible to those leaders for the results they achieve, but not beholden to the political system when making day-to-day decisions. 134. Some other principles and practices are particularly important for financial institutions such as a Pension Reserve Fund. These include • The purpose of the fund, such as the level and structure of the support to be provided, needs to be clearly defined. • Policies on accumulating the fund and for making withdrawals from it should be clear and made publicly available. • An annual report and accompanying financial statements should be prepared and made publicly available; the financial statements should be prepared in accordance with general international (or national, if appropriate) accounting standards and verified by an independent audit. • Financial transactions should be governed by a clear investment policy that addresses the objectives of the fund, the level of risk the fund is prepared to tolerate, the investment strategy it will follow, the role for external fund managers, the means used to select them, and the strategy to be employed to evaluate investment and investment manager performance. These policies and procedures should be publicly available. • The investment policy should also address any instances in which investment decisions are made for reasons other than to maximize risk-adjusted returns. These exceptions might include instance where investment reflect particular social or political concerns and should be disclosed publicly. • The fund will need to coordinate its activities with the fiscal and monetary authorities to make sure that fund actions do not interfere with the general management of the economy. -58- • The fund should prepare and disclose publicly the policies it will use to determine when and how it should exercise ownership rights associated with any investments it makes. C. Recommendations 135. A Pension Reserve Fund would be a useful mechanism to set aside some of the mineral royalties to support transitions to a more adequate and affordable pension system as well as finance a portion of future national pension costs. Setting aside such resources can satisfy social policy objectives of reducing the inter-generational reductions in pension replacement rates, improve the affordability of pensions in a rapidly aging society, and at the same time satisfy macro-economic objectives of spreading the distribution of the proceeds of mining royalties over time. 136. As suggested above, the view of this report is that such a Pension Reserve Fund is a preferable option to pre-finance future pension obligations when compared to establishing a Funded Defined Contribution design for the Tier II Pension Insurance Scheme. Such a Fund may potentially realize some returns in excess of wage growth while not subjecting participants to the risk of underperforming such a benchmark and at the same time potentially realize some of the effects of stimulating capital market development if accompanied by complementary legislation and institution building supporting the development and regulation of such markets. -59- VI. Governance and Institutional Reform 137. A detailed institutional assessment has not been undertaken and this section therefore does not suggest a detailed institutional reform program. Some areas have been identified however for institutional strengthening which will need to complement the policy reforms outlined above. 59 138. If the authorities decide to retain the NDC design then the policies and procedures for crediting notional accounts needs to be revised. The policies and procedures for the crediting of contributions and the notional interest rate need to be reviewed including strengthening internal and external audit procedures. The policies and procedures for accounting for service recognition prior to 1994 for post-1960 cohorts also need to be reviewed. 139. Regardless of the design decided upon for the Tier II pension insurance scheme, consideration will be needed for a buffer fund to manage the liquidity effects of demographic and economic shocks and to better align the inflows and outflows of Mongolia’s Tier II scheme as its population ages. In addition, some countries with NDC schemes have established automatic balancing mechanisms to ensure a sustainable balance between inflows and outflows in the face of changing demographic and economic circumstances. The authorities may want to consider such a mechanism. 140. Given the essential importance of public creditability to the functioning of the pension insurance scheme, several reforms are necessary in order to improve transparency and accountability. A substantial effort is needed to strengthen public credibility through web-based information and public information campaigns that inform the public of their entitlements and rights. Individuals need to be informed through web- access and annual statements both of their notional individual account balance and estimated benefits based on projected parameters to retirement age. Moreover, dispute resolution policies and procedures are needed to ensure that members can challenge any discrepancy identified in the crediting of contributions, notional interest, or debiting of disbursements and that these disputes are acted upon in a fair and judicious manner. Similarly, service standards and monitoring indicators to ensure that the SIGO is responsive to contributors and beneficiaries. These measures will aid in improving accountability and public credibility. 141. An essential part of ensuring the fairness of benefits under the NDC scheme is to have accurate data on life expectancy at retirement for covered workers. As post-1960 cohorts begin to retire, this data will have an important impact on benefit determination. 59 The governance and institutional arrangements needed for the establishment of a Pension Reserve Fund are indicated in Section V and Appendices 3 and 4. -60- Institutional capacity needs to be further developed so that credible mortality data can be produced and longevity monitored. -61- VII. Summary Recommendations and Possible Reform Process 142. The following is a summary of recommendations above: A. Objectives • Engage in a process to arrive at a policy consensus on the long-term target replacement rate from the Tier II Pension Insurance Scheme and overall pension instruments (Tier I-III), weighing the tradeoffs between adequacy of benefits and fiscal, employer and employee affordability. • Engage in a process to arrive at a policy consensus on the desired level of income and poverty protection for lower-income retirees through a targeted social pension. B. Overall Design • Consider a three tier design which includes a non-contributory Tier I Targeted Social Pension, a reformed Tier II pension insurance scheme, and introduction of Tier III supplementary savings arrangements to create a more diversified set of savings, insurance and redistributive instruments which can ensure old age income and poverty protection. C. Tier I – Targeted Social Pensions • Establish a Targeted Social Pension replacing the minimum pension guarantee for Tier II Pension Insurance and integrating the Social Welfare Pension. • Agree upon the benefit level, considering the impact on incentives to work, save and contribute to the Tier II scheme. 60 • Apply a pensions-test factor reduction based on the Pension Insurance benefit received by the retiree. Such a factor reduction should be scaled to ensure an incentive to contribute to the Tier II scheme. • Consider eligibility at age 60 (men and women) phased to age 65 at a rate consistent with the phasing of the increase in the retirement age for Tier II. D. Tier II – Contributory Pension Insurance scheme • As parametric changes to the defined-benefit scheme for pre-1960 cohorts will have a limited fiscal impact and could adversely affect public creditability, very limited changes have been suggested. • Remove the minimum pension guarantee and replace it with a Tier I Targeted Social Pension for both pre- and post-1960 cohorts on a phased basis. An initial level 60% of the MLS for Ulan Bataar, indexed to the growth in an elderly consumption basket 60 was proposed as one option. -62- • Gradually increase the retirement age to 65 for men and women (for post-1960 cohorts) and eliminate early retirement provisions so that workers have a stronger incentive to work longer. • Gradually reinstate a 19% contribution rate to improve benefit adequacy. • Adjust the annuity factor to be actuarially fair (if the NDC scheme is retained). • The benefit formulas for disability and survivorship need to be redesigned and could include flat insurance and wage-linked components. The benefits should be structured as pooled insurance products from earmarked contributions and the benefit formula aligned with periodic actuarial valuations. E. Tier III – Voluntary Occupational and Individual Pensions Savings Arrangements • Establish a funded-defined-contribution policy framework for voluntary occupational and individual pensions savings arrangements. • Establish a regulatory and supervisory framework for voluntary occupational and individual savings arrangements, including regulatory guidelines for investment management and the structuring of annuity arrangements. • Revise the tax treatment for licensed pensions and annuity arrangements. F. Addressing the Special Needs of Herders and the Informal Sector • Establish regulated individual retirement savings arrangements in accordance with the development of a Tier III supplementary voluntary pensions savings. • Elaborate a plan for matching defined contributions which would be provided for post-1960 cohorts participating in regulated voluntary savings arrangements. • Elaborate a plan for introduction of mandatory contributions to the largest herders, self-employed and informal sector workers. • Introduce a targeted social pension with the characteristics and features above. G. Pre-Financing Future Pension Liabilities • If the authorities decide to establish a Pension Reserve Fund, this could set aside some of the mineral royalties to help cover a portion of future national pension costs and facilitate transitions as Mongolia enacts revised pension policies. H. A Reform Process • Constitute a technical committee that would research and evaluate pension policy choices and advise policymakers on the advantages and disadvantages of difference design parameters and architectural choices. -63- • Develop a structured process of analysis and evaluation of policy and institutional reform options. Establish a mapping of this process including the sequencing of potential milestones (Figure 25). • Actively engage with multiple stakeholders to inform them of the rationale behind reform, the tradeoffs in adequacy and affordability for design choices, and build public awareness to strengthen public trust. -64- Figure 25. Potential Elements and Sequencing of a Policy Reform Process Timing Overall Reform Tier I Pension Tier II Tier III Pension Institutional Strategy for Policies and Strategy for Program Policies Pension Policies Arrangemen Pre-financing Incentives for Public Insurance ts Pension Increasing Worker Awareness & Policies Obligations and Coverage Building Public Transitions Trust Phase 1: • Constitution of a Enactment of a Enactment of Consideration of Tier Outlining Consideration of Evaluation of options Elaborate a 9 technical Tier I Targeted Tier II III Policy Reforms Institutional objectives, for matching public inf. months committee charged Social Pension Pension Reforms structuring, contributions and campaign with a work-plan legislation Insurance supporting governance & other incentives for explaining for investigation, Old Age the Tier I and investment voluntary current pension research and Reforms. Tier II strategy for a contributions to provisions, informing policy reform Pension Reserve pension savings and rationale for formulation. agenda Fund. insurance reform, & key • Est. a structured arrangements. merits of a select process for the set of reform technical options. committee to inform and advise policymakers. Phase 2: Continuation of the • Issuance of • Issuance of • Enactment of Tier III Revision of • Enactment of • Enactment of • Engagement 1 year work of the technical implement- old age Reforms, incl. a SIGO governing governing with the public committee including tation implement- regulatory operational legislation for legislation for a in how reform monitoring the guidelines, tation framework, a policies and a Pension Matching measures will results of reform. policy guidelines. supervisory body, procedures Reserve Fund. Contribution impact old age parameters • Enactment and tax treatment. to address • Est. of Scheme. income (benefit level, of Tier II • Institution building Tier I, Tier II governance, • Revision of security. indexation, disability of supervisory reforms + accountability applicability of Tier • Issuance of factor and capacity. Licensing management & II requirements for notional reduction). survivorshi of approved pension of an MDC transparency certain self- account p reforms. managers. scheme. policies and employed. statements and procedures, • Est. of revised projected investment compliance entitlements. policies, enforcement guidelines. Phase 3: “ Calibration of • Issuance of Elaboration of Put the Pension Gradual scaling up of Est. of a 2 years benefit level, disability product guidelines Reserve Fund enforcement retirement indexation, and for occupational and into operation guidelines with literacy factor reduction survivor- individual retirement incl. investment increasing campaign and based on ship savings management, applicability to self- information implement- implement. arrangements. disclosure and employed. toolkits for tation guidelines. review. retirement experience. planning. -65- VIII. Conclusions 143. This report is intended as a starting point for consideration of reform designs and parameters noting that additional work is needed to refine the core objectives, designs and reform parameters. It is intended to assist the authorities in considering a comprehensive reform design by outlining the key design options and parameters for further consideration. 144. The report identifies several needs for reform of pension and elderly assistance in Mongolia’s rapidly changing economy and aging society. It suggests that anticipated reductions in retirement benefits beginning in 2015 could serve as a basis for a more comprehensive review and reform of the pension insurance scheme and elderly social assistance programs. It also suggests that the recent growth of the mining sector along with growth in volatility present both challenges and opportunities. The challenges are to reform the provision of pensions to improve adequacy and affordability while at the same time limiting the long-term requirements for fiscal support. Furthermore, the pension system needs to be able to adjust to anticipated wage and other economic volatility and at the same time adapt to a dramatically aging society. The opportunities presented by the growth of mineral proceeds are that a portion of such funds could assist in supporting a transition program to set Mongolia upon a transition path towards achieving the reform design and parameters considered. 145. To address these challenges, the report suggests that the authorities review the societal objectives of the combined public and private pension provision considering: (i) the target level of income replacement in old age; (ii) the desired and achievable level of affordability for the Treasury, employers and employees; and (iii) the balance between work-life and retirement as manifest in a policy decision on the age of becoming entitled to pension benefits. 146. A three tier design is proposed as a means of strengthening old age income protection. The proposal largely retains the design characteristics of the current pension insurance but modifies provisions for retirement age, benefit indexation, minimum pension provisions, as well as the design and financing of the disability and survivorship benefits. The proposed architecture also includes a targeted social pension, which would provide for basic subsistence for the elderly in the bottom half of the retiree income distribution which should also replace the current minimum pension guarantee. Parametric reforms proposed for the pension insurance scheme can reduce the required state subsidy a portion of which can then defray the anticipated growth in costs for an expanded targeted social pension. -66- 147. The report also considers options for pre-financing transition arrangements and future pension costs through the establishment of a Pension Reserve Fund constituted from a portion of mining royalties. Setting aside such resources can satisfy social policy objectives of reducing the inter-generational reductions in pension replacement rates, improve the affordability of pensions in a rapidly aging society, and at the same time satisfy macro-economic objectives of spreading the distribution of the proceeds of mining royalties over time. -67- Bibliography Asian Development Bank, 2006. Technical Assistance for Strengthening the Pension System (Cofinanced by the Government of Luxembourg). December 20, 2006. Erdene, Uuganbileg, 2004. Social Assistance Program Assessment for Mongolia, Draft, memo, April 2004. Fox, Louise, 2002. Mongolia: Cash Transfers and Social Protection Programs, World Bank, draft mimeo, March 25, 2002. Ghazy, Mujahid; Oyut-Erdene Namdaldagva, Oyun Banzragch, 2010. Assuring Income Security in Old Age: Views of the Mongolian Elderly, Ulaanbaatar. Government of Mongolia, 2003. Ministry of Social Welfare and Labour, Social Security Sector Strategy Paper, Ulaanbaatar, November. Government of Mongolia, 2001. State Social Insurance General Office, Social Insurance Legislation, Ulaanbaatar. Government of Mongolia, Ministry of Social Welfare and Labour, 2010. Facts and Figures 2010, Ulaanbaatar. Government of Mongolia, State Social Insurance General Office, 2006, 2007. The Statistical Yearbook of Social Insurance, Ulaanbaatar, 2006, 2007. Government of Mongolia, State Great Hural, 2009. Resolution 39 of the State Great Hural, Ulaanbaatar. Holzmann, Robert and Hinz, Richard, 2006. Old Age Income Support in the 21st Century, World Bank. http://siteresources.worldbank.org/INTPENSIONS/Resources/Old_Age_Income_Support_FM.p df Holzmann, Robert and Palmer, Edward, 2006. Pension Reform: Issues and Prospects for Non- Financial Defined Contribution (NDC) Schemes, World Bank. http://go.worldbank.org/K9HH10CS70 Holzmann, Robert, Robalino, David, Takayashi, Noriyuki, 2009. Closing the Coverage Gap: The Role of Social Pensions and Other Retirement Transfers, World Bank. http://siteresources.worldbank.org/INTPENSIONS/Resources/Closing_Coverage_Gap.pdf International Labor Organization, 2010, Mongolia Social Security Reform (Draft). IMF, 2011, Public Information Notice on Second Post –Program Monitoring, Public Information Notice (PIN) No. 11/146, November 28, 2011. IMF, 2011, Mongolia: 2011 Article IV Consultation—Staff Report; Staff Supplement. http://www.imf.org/external/pubs/ft/scr/2011/cr1176.pdf -68- IMF, 2010, Mongolia: Joint IMF/World Bank Debt Sustainability Analysis Under the Debt Sustainability Framework for Low-Income Countries. http://www.imf.org/external/pubs/ft/scr/2010/cr10166.pdf IMF, Fiscal Affairs Department, 2009. Mongolia: Expenditure Rationalization in the Context of Falling Mineral Revenue, January 15, 2009. IMF Fiscal Affairs Department, 2005. Options for Expenditure Savings and Efficiency Improvements, January 2005. International Working Group of Sovereign Wealth Funds, 2008. “Sovereign Wealth Funds: Generally Accepted Principles and Practices – ‘Santiago Principles’.� October . OECD, 2011. Pensions at a Glance, 2011. OECD Paris. OECD, 2011. Pensions at a Glance, Asia. OECD, Paris. OECD, 2010. Sickness, Disability and Work: Breaking the Barriers - A Synthesis of Findings across OECD Countries. http://www.oecd-ilibrary.org/social-issues-migration-health/sickness- disability-and-work-breaking-the-barriers_9789264088856-en OECD, 2009, Pensions at a Glance, 2009. OECD Paris. Palacios, Robert and Whitehouse, Edward, 2006, Civil Service Pension Schemes around the World, World Bank. http://siteresources.worldbank.org/INTPENSIONS/Resources/Old_Age_Income_Support_FM.p df State Insurance and General Office, 2010. SIGO Yearbook 2010, Social Security Development Programme Project and Ministry of Social Welfare and Labour, 2006, Expansion of Pension Insurance Coverage & Reform of Insurance System, Study Report, Ulaanbaatar. Thompson, Lawrence H., 2003. Analysis of Social Security Policy Options, Report of the Working Group on Social Insurance, Ministry of Social Welfare and Labor of Mongolia, January 17, 2003. Wiese, Patrick, 2006. An Evaluation of Pension Reform Options for Mongolia, December 25, 2006. World Bank. 2010. Mongolia Quarterly Economic Update, October 2010. http://siteresources.worldbank.org/INTMONGOLIA/Resources/Mongolia_Quarterly_Economi c_Update_Oct_2010.pdf World Bank. 2009. Mongolia Consolidating the Gains, Managing Booms and Busts, and Moving to Better Service Delivery. A Public Expenditure and Financial Management Review. http://www- wds.worldbank.org/external/default/WDSContentServer/WDSP/IB/2009/02/08/000333037 _20090208230004/Rendered/PDF/433530ESW0MN0v1eb05020090Box334131B.pdf World Bank. 2008. Mongolia: Pension Policy Challenges and Reform Options, Human Development Sector Unit, East Asia and Pacific Region, World Bank, Washington D.C. -69- World Bank. 2007. Mongolia: Building the Skills for the New Economy, Report No. 40118, Human Development Sector Unit, East Asia and Pacific Region, World Bank, Washington D.C. http://siteresources.worldbank.org/INTMONGOLIA/Resources/building_the_skills_for_new_ec onomy_ENG.pdf World Bank. 2006. Mongolia: Promoting investment and job creation: An investment climate assessment and trade integration study, World Bank, Washington D.C. -70- Appendix 1: Social Insurance Pension Parameters61 Defined-Benefit Scheme (pre-1960 Cohorts) NDC Scheme (post 1960 cohorts) Old Age Insurance Law on Social Insurance and Law on Pensions and Benefits Individual Pension Insurance Provided by the Fund of Social Insurance 1994 (as amended) Contribution Accounts Law of 1999 Applicability All contract employees born after January 1, 1960. Contribution Rate 14% of wages (7.0% employer and 7.0% employee) 14% of wages Minimum Wage National Minimum Wage (revised periodically) National Minimum Wage (revised subject to periodically) Contributions Maximum Covered None None Wage Benefits - Accrual 2.25%/year for the first 20 years of eligible service and 1.5%/year Based on Notional Account balance for rate for each year of eligible service after that. 15% contribution rate for years of contributions, accrued notional returns for each year (average growth in the last three years’ average wages), and average life expectancy factor. Wage base for Best 5 years’ consecutive wages out of the final 20 years wages Not applicable. benefit reported. determination Minimum Pension 75% of the minimum wage. 20 percent of the national average wage, plus an additional 0.5 percent of the average wage for each additional service year beyond the minimum of 15 years Indexation In practice, pension increases are determined by Parliament on an “in direct relation to the inflation rate� ad-hoc basis, although the law states that pensions should be The pension adjustment index and the increased in “relation to changes in the cost of living�. procedure for its application shall be determined by the National Statistical Office based on the suggestion of the Social Insurance National Council. Taxation of benefits Tax exempt Tax exempt Qualifying Conditions Vesting • 20 years • 15 years of service and contributions Requirements Retirement Age • 60 for men and 55 for women not subject to special early same retirement rules. • 55 for men with 20 years of service, 12 of which in hazardous conditions. • 50 for women with 20 years of service, 10 of which in hazardous conditions. • 50 for men with 20 years of service, 10 of which underground or in high heat. • 45 for women with 20 years of service, 7 of which underground or high heat. • 50 for women who have brought up 4 or children. Invalidity/Disability Enabling Law On Pensions & Benefits Provided By The Social Insurance Law On Individual Pension Insurance Legislation Fund Enforced on 1/1/1995, Amended in 1995, 1996,1997, 1999, Contribution Account. Effective 2002, 2003, 2004, 2006, 2008 and 2009 7/1/1999 Contributions Contained within the combined 14% contribution rate. Contributions for invalidity, survivorship, minimum pension guarantees and administrative costs all contained in 14% contribution. 61 This table has been reprinted from State Social Insurance General Office, Social Insurance Legislation, 2001, in Louse Fox, Cash Transfers and Social Protection Programs, 2002 and modified to reflect changes legislated in 2008. -71- Defined-Benefit Scheme (pre-1960 Cohorts) NDC Scheme (post 1960 cohorts) Benefit Formula DP = W*45% * percentage of capacity loss(I) Note: This law is not applicable to DP = W* 45% * percentage of capacity loss(II) injury related benefits. DP = Disability pension W = Insured person’s monthly average wage PCL(I) = over 50% PCL(II) = 100% if lost 70 percent of working ability Formula Total invalidity = same formula as retirement Total invalidity = W*60% Partial Invalidity = (retirement formula)* percentage of loss of Partial invalidity = (W*.6) * capacity for work (percentage of loss of capacity for work) W = monthly average wage in the last three years, Length of pension From the date of commencing invalidity and ending with the day Same as under the Current Law of rehabilitation, or with the following month in which the beneficiary dies Minimum pension Same as the minimum retirement pension Same as the minimum retirement pension Qualification Not less than 50 per cent loss of capacity for work permanently, or Same for a long duration due non-occupational disease or injury Classification of Total invalidity (70% and more loss of working capacity) and Same invalidity partial invalidity (from 50 to70%) Minimum length of Not less than 20 years insured service, or three years out of five, Same service for immediately preceding the date of commencement of invalidity qualification Incomplete service Disability benefit is reduced proportionally to the period of Not allowed (minimum 3 years) pension insurance, but its minimum is equal to minimum reduced retirement pension Minimum pension Not less than 75 per cent of the wage in case of working capacity loss over 30% and more Length of pension From the date of commencing disability and ending with the day of rehabilitation, or with the following month in which the beneficiary dies Certification Determined by Medical Labor Accredited Commission, whether Same cases are relevant to employment injury determined by Employer’s Investigating & Registering Commissions. Dependent’s/Survivorship Pension Applicable Law on Benefits Provided By The Fund Of Social Insurance Law On Individual Pension Insurance Legislation Against Employment Injury & Occupational Diseases. Enforced Contribution Account on 1/1/1995, Amended in January 2000 Effective 7/1/1999 Benefit formula Same benefit formula as old age retirement; 40% monthly average wage in the last Survivors receive pension in proportion to certain percentages three years for one dependent increased depending upon the number of survivors: for three or more -100%, by 10% per each member over two and two - 75% and one - 50%. more. But pension should not exceed 60% monthly average wage Eligibility Family dependents determined as survivors under the Law On Pensions & Benefits Provided By The Social Insurance Fund Length of pension Same as survivor’s pension under the Law On Pensions & Benefits Provided By The Social Insurance Fund Credit for period of Pension increased by 1 per cent pension amount for each year of No credit invalidity pension total invalidity, if the deceased was on the receipt of invalidity pension Minimum pension 50% minimum wage for one dependent, 75% for two dependents Same as the minimum retirement and 100% for three dependents; reduced pension should not be pension provided from Individual less than 50% same wage. Account Defined-Benefit Scheme (pre-1960 Cohorts) NDC Scheme (post 1960 cohorts) Length of service Not less than 20 years insured service, or three years out of five, Same immediately preceding the date of breadwinner’s death due to non-occupational disease or accident. Incomplete service Pension is reduced proportionally to the period of pension Not allowed (minimum 5 years; insurance. -72- continuous in last year) Eligible survivors Born or adopted children under 16, if student children to age 19; Same grandchildren, brothers and sisters under 16 without caregivers; grandchildren, brothers and sisters disabled or got disabled prior attaining age 16; parents over retirement age or disabled parents, spouse or grandparents, brothers and sisters without caregivers; any of parents or spouse not working and caring for children under 8, or grandchildren and younger brothers and sisters, also family dependents of the deceased who was on receipt of retirement or invalidity pension and who totally lost capacity for work in months preceding the death; step-parents; step-children not receiving alimony from their own parents; family dependents of the disappeared; Length of pension From the date of death up to the day on which surviving children Same attain age 16, if student children to age 19, incapacitated persons for period of loss of working capacities, survivors who have attained the pension age up to the end of the following month in which the insured breadwinner’s dies Source: State Social Insurance General Office, Social Insurance Legislation, 2001, in Louise Fox, Cash Transfers and Social Protection Programs, 2002. -73- Appendix 2: Glossary62 Accrual rate. The rate at which pension entitlement is built up relative to earnings per year of service in earnings-related schemes—for example, one-sixtieth of final salary per year of applicable service. Accrued pension. The value of the pension to a member at any point prior to retirement, which can be calculated on the basis of current earnings or also include projections of future increases in earnings. Actuarial fairness. A method of setting insurance premiums according to the true risks involved. Additional voluntary contributions. Contributions to an occupational pension scheme over and above the employee’s normal contribution rate. Adverse selection. A problem stemming from an insurer’s inability to distinguish between high- and low-risk individuals. The price for insurance then reflects the average risk level, which leads low-risk individuals to opt out and drives the price of insurance still higher until insurance markets break down. Annuity. A stream of payments at a specified rate, which may have some provision for inflation proofing, payable until some contingency occurs, usually the death of the beneficiary or a surviving dependent. Annuity factor. The net present value of a stream of pension or annuity benefits. Annuity rate. The value of the annuity payment relative to its lump-sum cost. Average effective retirement age. The actual average retirement age, taking into account early retirement and special regimes. Contribution ceiling. A limit on the amount of earnings subject to contributions Commutation. Exchange of part of the annuity component of a pension for an immediate lump sum. Contracting out. The right of employers or employees to use private pension fund managers instead of participating in the publicly managed scheme. Current balance. The current balance is total contributions, revenues and other income accrued in a particular period minus all disbursements and other expenditures. Deferred annuity. A stream of benefits commencing at some future date. Defined benefit. A pension plan with a guarantee by the insurer or pension agency that a benefit based on a prescribed formula will be paid. Can be fully funded or unfunded and notional. Defined contribution. A pension plan in which the periodic contribution is prescribed and the benefit depends on the contribution plus the investment return. Can be fully funded or notional and nonfinancial. Demogrant. Same as a universal flat benefit, where individuals receive an amount of money based solely on age and residency. Demographic transition. The historical process of changing demographic structure that takes place as fertility and mortality rates decline, resulting in an increasing ratio of older to younger persons. Disclosure. Statutory regulations requiring the communication of information regarding pension schemes, funds, and benefits to pensioners and employees. 62 This glossary is drawn from Robert Holzmann and Richard Hinz, Old Age Income Support in the 21st Century, The World Bank, 2006. -74- Discretionary increase. An increase in a pension payment not specified by the pension scheme rules. Early leaver. A person who leaves an occupational pension scheme without receiving an immediate benefit. Early retirement. Retirement before reaching an occupational scheme’s normal retirement age or, in the state scheme, before reaching the state’s pensionable age. Earnings cap (ceiling). A limit on the amount of earnings subject to contributions. Financing gap. Present value of the current balances projected into the future. Full funding. The accumulation of pension reserves that total 100 percent of the present value of all pension liabilities owed to current members. Funded defined contribution scheme. A scheme in which contributions and investment returns are held in an actively and competitively managed fund and the returns realized on the assets (after all fees and administrative costs) are credited to an individual’s account balance. Funding. Accumulation of assets in advance to meet future pension liabilities. Implicit pension debt (net). The value of outstanding pension claims on the public sector minus accumulated pension reserves. Indexation. Increases in benefits by reference to an index, usually of prices, although in some cases of average earnings. Inter-generational distribution. Income transfers between different age cohorts of persons. Intra-generational distribution. Income transfers within a certain age cohort of persons. Legal retirement age. The normal retirement age written into pension statutes. Marginal pension. The change in the accrued pension between two periods. Means-tested benefit. A benefit that is paid only if the recipient’s income falls below a certain level. Minimum pension guarantee. A guarantee provided by the government to bring pensions to some minimum level, possibly by “topping up� the capital accumulation needed to fund the pensions. Moral hazard. A situation in which insured people do not protect themselves from risk as much as they would have if they were not insured. For example, in the case of old- age risk, people might not save sufficiently for themselves if they expect the public system to come to their aid. Nonfinancial (or notional) defined-contribution (plan). A defined-benefit pension plan that mimics the structure of (funded) defined-contribution plans but remains unfunded (except for a potential reserve fund). Normal retirement age. The usual age at which employees become eligible for occupational pension benefits, excluding early-retirement provisions. Notional (or nonfinancial) accounts. Individual accounts where the notional contributions plus interest rates accrued are credited and determine the notional capital (that is, the liability to society). Notional (or nonfinancial) capital. The value of an individual account at a given moment that determines the value of annuity at retirement or the transfer value in case of mobility to another scheme or country. Notional or nonfinancial interest rate. The rate at which the notional accounts of notional defined-contribution plans are annually credited. It should be consistent with the financial sustainability of the unfunded scheme (potentially the growth rate of the contribution base). -75- Occupational pension scheme. An arrangement by which an employer provides retirement benefits to employees. Old-age dependency ratio. The ratio of older persons to working-age individuals. The old- age dependency ratio may refer to the number of persons over 60 divided by, for example, the number of persons ages 15–59, the number of persons over 60 divided by the number of persons ages 20–59, and so forth. Overannuitization. A situation in which a compulsory pension forces an individual to save more in pension than he or she would in the absence of the compulsory provision. Pay-as-you-go. In its strictest sense, a method of financing whereby current outlays on pension benefits are paid out of current revenues from an earmarked tax, often a payroll tax. Pension coverage rate. The number of workers actively contributing to a publicly mandated contributory or retirement scheme, divided by the estimated labor force or by the working-age population. Pension lump sum. A cash withdrawal from a pension plan, which in the case of some occupational pension schemes is provided in addition to an annuity. Also available from personal pension plans. Pension spending. Usually defined as old-age retirement, survivor, death, and invalidity- disability payments based on past contribution records plus noncontributory, flat universal, or means-tested programs specifically targeting the old. Pensionable earnings. The portion of remuneration on which pension benefits and contributions are calculated. Portability. The ability to transfer accrued pension rights between plans. Replacement rate. The value of a pension as a proportion of a worker’s wage during a base period, such as the last year or two before retirement or the entire lifetime average wage. Also denotes the average pension of a group of pensioners as a proportion of the average wage of the group. Supplementary pensions. Pension provision beyond the basic state pension on a voluntary basis. Support ratio. The opposite of the system dependency ratio: the number of workers required to support each pensioner. System dependency ratio. The ratio of persons receiving pensions from a certain pension scheme divided by the number of workers contributing to the same scheme in the same period. System maturation. The process by which a pension system moves from being immature, with young workers contributing to the system, but with few benefits being paid out since the initial elderly have not contributed and thus are not eligible for benefits, to being mature, with the proportion of elderly receiving pensions relatively equivalent to their proportion of the population. Transition costs. Costs of moving from one pension policy design to another. Generally used to refer to costs of moving from a pay-as-you-go financing approach to a fully funded approach. Universal flat benefit. Pensions paid solely on the basis of age and citizenship, without regard to work or contribution records. Valorization of earnings. A method of revaluing earnings by predetermined factors such as total or average wage growth to adjust for changes in prices, wage levels, or economic growth. In pay-as-you-go systems, pensions are usually based on some -76- percentage of average wages. This average wage is calculated over some period of time, ranging from full-career average to last salary. If the period for which earnings history enters into the benefit formula is longer than the last salary, the actual wages earned are usually revalued to adjust for these types of changes. Vesting period. The minimum amount of time required to qualify for full and irrevocable ownership of pension benefits. -77- Appendix 3: Options for Pre-Financing Pension Liabilities The Mongolian government has set aside a portion of its mineral royalties in a human development fund and another portion in a fiscal stabilization fund; it is considering creating pension and commodity funds in the future. The human development fund is currently disbursing cash payments to each citizen and also promises each citizen education and health benefits. The fiscal stabilization fund gives the government fiscal cushion to deal with unforeseen, adverse fiscal developments. The current plan is to deposit excess mineral royalty revenue into the fund until the balance reaches a minimum of 5 percent and a maximum of 10 percent of GDP. Ministry of Finance officials expect the maximum to be reached around 2015. That is the point at which they may decided to establish a pension fund and/or one or more commodity stabilization funds. A pension fund would be a mechanism to allow setting aside some of the mineral royalties to help cover a portion of future national pension costs. Similar actions are occurring in a number of countries around the world. Among OECD countries, for example, pension investment funds have been established in Japan, Korea, Canada, Norway, Ireland, and New Zealand. Each country anticipates increases in their national pension plan’s costs and seeks to accumulate reserves now to help finance these future costs. None anticipate fully-funding future pension liabilities; their objective is simply to cover a portion of the projected increase from investment income, reducing the impact of the cost increase on future generations. Although the fiscal stabilization fund does not seem to be controversial, the human development fund has been. In particular, some question the wisdom of making cash payments to individuals and others question the particular strategy adopted to support higher education. The next year or two might, therefore, be a good time for Mongolian parliamentarians and other government officials to see if they can agree on a general strategy for the use of the projected mineral royalties after the human development fund expires. An agreement on a general strategy is important because of the variety of different ways that the mineral royalties might be used. In the absence of a clear plan for their use, Mongolia runs the risk of letting shifting short-term political currents cause it to lurch from one strategy to another, never fully debating the relative merits of each and probably never fully achieving the potential benefits of any. In additional to the current actions and currently proposed uses, royalty income might be used on an ongoing basis for other investments in human capital, investments in physical infrastructure, helping to finance a new targeted pension , supporting the current contributory pension system, or simply as one source of general government revenue. Moreover, while royalties could be accumulated in a fund designed specifically to deal with future pension costs, they could also be set aside to deal with future balance of payments problems, used to cushion the impact of future fluctuations in world commodity prices, or simply saved for unspecified future national needs to be determined later (i.e., a sovereign wealth fund) . -78- The potential uses vary widely in their objectives, their mechanics and their likely impact. Some involve support for current public or private consumption, some support current spending on various investment projects. Some seek to accumulate funds to deal with future contingencies that are likely to occur but whose scope and timing can’t be known. Others seek to accumulate funds for liabilities (such as pension costs) that can be projected with a fair degree of certainty. The variety of uses and the differences among them in implied operating principles and procedures is the reason that a political consensus is an important next step. A general strategy for the use of the royalties may amount to little more than a general agreement as to the likely flow to the government of royalty revenues over the next 20 to 30 years and a rough plan for allocating those resources among the competing uses. Obviously, any plan adopted through such a process would be modified over the years as new information emerged and relative priorities changed, but the plan would provide a structure for future discussions and increase the odds that the royalties will be used wisely in pursuit of long-term objectives. Discussion of how royalties might be allocated among these competing uses should take into consideration the foreign exchange implications of the exploitation of Mongolia’s mineral resources. Other things equal, a major increase in mineral exports would cause the value of the Tugrik to rise relative to other world currencies. In the first instance, this will benefit Mongolian consumers by making imported goods cheaper. But it could harm other sectors of Mongolian industry and the people who work in other sectors by making it more difficult to export other domestically produced products. Of the competing uses for royalty income, some uses could be structured to address this problem by limiting the appreciation of the Tugrik, while others would have no direct effect on its value. In particular, pension funds, commodity price and balance of payment stabilization funds or sovereign wealth funds can be structured to help limit current appreciation by investing foreign exchange reserves abroad. Indeed, a fund designed to help stabilize the balance of payments must be held in instruments denominated in a foreign currency, and pension and sovereign wealth funds ought to hold substantial investments outside of Mongolia. 63 In contrast, using royalty income for direct payments to Mongolian citizens, direct investments in human or physical capital, or general budget support will not help deal with the problem of currency appreciation. Weighing the likely size of and possible consequences for the domestic economy of currency appreciation will be an important factor to consider in deciding how to use the royalty income. If it is decided to save a portion of the royalty income through one or more special investment funds, it is important to be clear about the precise purpose of each of fund because a fund’s operations will differ according to what its objective is. For example, assets held to help finance future pension liabilities will probably not be needed for several 63 Withdrawals from a pension fund can reduce the fiscal burden of financing pension benefits, but if they are achieved by selling domestic assets to other Mongolian citizens, they will not reduce the economic burden on society of paying those benefits. On the other hand, assets invested abroad, particularly in countries from which Mongolia would ordinarily import consumer goods, would allow the payment of pension benefits without imposing a burden on the domestic economy because they could be used indirectly to finance consumption imports. -79- decades. They can be invested in less liquid instruments and in instruments with longer durations and, presumably, higher average returns. On the other hand, assets held in a stabilization fund need to be available on short notice and must therefore be invested in shorter duration, liquid instruments. If it is decided to pursue both the objective of price or balance of payment stabilization and the objective of helping to finance future pension liabilities, two (or more) separate funds should be created. Trying to do both through the same fund is likely to force the fund managers to make investment decisions that are not optimal for either objective. Setting up a national pension fund involves critical decisions in at least three different areas: (1) funding and withdrawal rules, (2) institutional structure, and (3) investment policies. 64 Funding and withdrawal rules involve determining the schedule for deposits into the fund and the conditions under which fund assets will be disbursed. Presumably, a tentative schedule for depositing royalty income into the fund will be developed based on consideration of the expected total royalty income available, the competing uses for that income, the foreign exchange consequences of different approaches, and the likely impact of different accumulation schedules on the fund’s ability to offset future pension costs. The accumulation schedule is likely to be a plan that is subject to annual review and adjustment as more is known about actual royalty income and the value of each alternative use. At least initially, the withdrawal plan may involve the prohibition of any withdrawals until a specified future date or until the fund reaches a certain size. Thereafter, it may involve a specified goal as to the fraction of future costs to be met from the fund. 65 For example, in New Zealand, deposits into the pension fund come from general government resources and are supposed to be made according to an actuarially- determined schedule. Annual deposits average about 0.75 percent of GDP. It is expected that withdrawals will begin around 2030 and will support 15% to 20% of the cost of the New Zealand superannuation scheme. The objective of the Canadian pension fund is to supply sufficient resources to keep future contribution rates from rising above a specific target level. As would be the case in Mongolia, Norway operates a fund that receives income from mineral royalties. Although started in 1990 as the Norwegian Petroleum Fund, it was redesignated a pension fund in 2006. Withdrawals are limited to four percent of its capital in any given year on the assumption that on average the fund will earn a four percent annual real rate of return. 66 64 See, for instance, Udaibir S. Das, Yinqiu Lu, Christian Mulder, and Amadou Sy, “Setting up a Sovereign Wealth Fund: Some Policy and Operational Considerations,� IMF Working Paper 09/179, and Cornelia Hammer, Peter Kunzel, and Iva Petrova, “Sovereign Wealth Funds: Current Institutional and Operational Practices,� IMF Working Paper 08/254. 65 The relationship between the size of a pension fund and the level of expenditure support it can supply on an ongoing basis is given by the equation: g = (r-e)*R, where g is the level of spending that can be supported from the fund, as a percent of GDP, r is the average rate of return earned on the fund, e is the average rate of growth in total pension expenditures and R is the ratio of the balance in the fund to GDP. Thus if r is 5 percent, e is 2 percent and the balance in the fund is 50 percent of GDP, then the fund can support expenditures totaling 1.5 percent of GDP on an ongoing basis. 66 Juan Yermo, “Governance and Investment of Public Pension Reserve Funds In Selected OECD Countries� Organization for Economic Cooperation and Development, Financial Market Trends, No. 94, Vol 2008/1. -80- A variety of institutional arrangements are used around the world to govern and manage national pension funds. In some cases, assets are managed by the same agency that also manages the pension plan; in other cases, asset management responsibilities are lodged in a separate institution. A number of pension plans in the developing world either operate as provident funds or have reserves that were accumulated in the plan’s early years when benefit payments were low. In some, such as the Employee Provident Fund in Malaysia and the Social Security System in the Philippines, assets are managed directly by the pension agency. In others, such as the Central Provident Fund of Singapore, assets are managed by a separate institution. As a general rule, national pension assets held in developed countries are also managed by separate agencies. There is no hard evidence as to the superiority of one institutional arrangement over another, but international expert opinion probably favors lodging asset management responsibilities in a separate agency for two main reasons. First, asset management requires a set of skills and operating modes entirely different from the skills and operating modes needed to run an effective pension agency. As there is little synergy between the two kinds of activities, there is little to be gained by housing them in the same institution. Second, it is likely that the asset management function can be better insulated from day to day political interference if is lodged in a separate institution that is designed to be somewhat independent. In some cases, particularly when the fund is just beginning and/or not expected to ever be very large, the asset management function can be lodged at the central bank. Where the size of the fund justifies the expense of creating a separate organization, these are usually created with boards of directors composed of individuals who are appointed to fixed terms by the political leaders, responsible to those leaders for the results they achieve, but not beholden to the political system when making day-to-day decisions. A second issue involves the use of outside managers to make and execute actual investment decisions. Again, particularly among developed countries, the trend is to use outside managers for this function. In Mongolia’s case, these would likely be firms located in one or more of the Asian financial centers (Hong Kong, Tokyo, or Singapore). The managers of the Mongolian fund would select the asset managers, work with them to develop the investment strategy to be used, and agree on metrics to be used in assessing their performance. But the Mongolian officials would not necessarily be involved selecting particular investment, executing trades or taking custody of the actual financial assets. Regardless of the particular institutional structure selected, operating policies and procedures will have to be developed to limit risk and assure transparency. Several recent papers discuss the breadth of the concerns that need to be addressed. 67 In addition, a group representing the world’s major sovereign wealth funds has agreed to a set of generally accepted principles and practices for such funds, the “Santiago Principles,� which also would be applicable to a national pension fund. 68 67 For example, see Organization for Economic Cooperation and Development, “Guidelines for Pension Fund Asset Management,� adopted by the OECD Council January 26, 2006. Also, see Carmichael, Jeffrey and Robert J. Palacios, “Managing Public Pension Funds: A Framework.� Paper presented to 2nd Public Pension Fund Management Conference. World Bank, Washington DC, May 2003. 68 International Working Group of Sovereign Wealth Funds, “Sovereign Wealth Funds: Generally Accepted Principles and Practices – ‘Santiago Principles’.� October 2008. The working group was set up by the International -81- Many of the generally accepted principles articulated by the group of sovereign wealth funds represent organizational and management practices appropriate in a wide range of government and private sector organizations. They include the need for an organizational structure that produces clear and effective division of responsibilities, assures adequate authority for executing responsibilities and a clear system of accountability for reviewing actions. They also call for the timely reporting of all relevant statistical information to relevant authorities, promulgation of clear standards of professional and ethical conduct, and the maintenance of a reliable structure of internal controls for effective risk management. Some other principles and practices are particularly important for financial institutions such as a national pension fund. These include: • The purpose of the fund, such as the level and structure of the support to be provided, needs to be clearly defined. • Policies on accumulating the fund and for making withdrawals from it should be clear and made publicly available. • An annual report and accompanying financial statements should be prepared and made publicly available; the financial statements should be prepared in accordance with general international (or national, if appropriate) accounting standards and verified by an independent audit. • Financial transactions should be governed by a clear investment policy that addresses the objectives of the fund, the level of risk the fund is prepared to tolerate, the investment strategy it will follow, the role for external fund managers, the means used to select them, and the strategy to be employed to evaluate investment and investment manager performance. These policies and procedures should be publicly available. • The investment policy should also address any instances in which investment decisions are made for reasons other than to maximize risk-adjusted returns. These exceptions might include instance where investment reflect particular social or political concerns and should be disclosed publicly. • The fund will need to coordinate its activities with the fiscal and monetary authorities to make sure that fund actions do not interfere with the general management of the economy. • The fund should prepare and disclose publicly the policies it will use to determine when and how it should exercise ownership rights associated with any investments it makes. Clearly, establishing a pension fund requires the authorities to make a number of policy and operational decisions. Fortunately, a number of other organizations around the world have experience in setting up similar kinds of institutions and in making similar policy and operational decisions. It should not be difficult for Mongolia to get adequate assistance if it should elect to create such a fund. The first step, however, is the development of a political consensus as to the desirability of such a fund, its approximate size and its general purpose. Monetary Fund and has a web site at www.iwg-swf.org . The 24 Generally Accepted Principles and Practices are summarized in Appendix 4. -82- Appendix 4: Summary of Generally Accepted Principles and Practices (GAPP) – the “Santiago Principles�69 1. The fund should have a clear and effective legal framework that supports its objectives and ensures legal support for its transactions. Key features of the fund’s legal basis and structure and its relationship with other government institutions should be publicly disclosed. 2. The fund’s policy purpose should be clearly defined and publicly disclosed. 3. Fund activities should be closely coordinated with the monetary and fiscal authorities to assure consistent macroeconomic policies. 4. The fund needs clear (and publicly disclosed) policies, rules and procedures relating to the size and source of funding and the subsequent disbursement of the fund’s assets. 5. Relevant statistical data describing the fund’s operations need to be reported to relevant government authorities for inclusion in macroeconomic data sets. 6. The fund needs a sound governance framework that includes a clear and effective division of roles and responsibilities to facilitate accountability and operational independence. 7. The government should set the fund’s objectives, appoint members of its governing body, and monitor its performance in accordance with clearly defined procedures. 8. The governing body should have a clear mandate and adequate authority to act in the best interests of the fund. 9. Fund managers should implement the fund’s strategies in accordance with clearly defined responsibilities. 10. A framework for assuring accountability in the funds operations should be clearly defined in the relevant legislation, charter or other documents establishing the fund’s basic structure. 11. An annual report and accompanying financial statements should be prepared following international (or, if applicable, national) accounting standards. 12. The fund’s operations and financial statements should be audited annually in accordance with generally recognized audit standards. 13. Clear professional and ethical standards should be established and communicated to members of the governing body, as well as the funds management and staff. 69 International Working Group of Sovereign Wealth Funds, “Sovereign Wealth Funds: Generally Accepted Principles and Practices – ‘Santiago Principles’.� October 2008. The working group was set up by the International Monetary Fund and has a web site at iwg-swf.org. -83- 14. All dealings with third parties should follow clear rules and procedures and be based on the economic and financial interests of the fund. 15. Operations and activities must comply with applicable national regulatory and disclosure requirements. 16. Both the fund’s governance framework and objectives and the manner in which its management is independent of the government should be publicly disclosed. 17. Information needed to contribute to stability in international financial markets and enhance confidence among other countries in the fund’s integrity should be publicly disclosed. 18. The fund should have a clear and consistent investment policy, defining the fund’s objectives, its risk tolerance (including possible use of leverage) and investment strategy. The investment policy should also address the use of internal and external investment managers, the range of their authorities and activities, the process for their selection, and the procedure for monitoring their performance. This policy should be disclosed publicly. 19. Investment decisions should aim to maximize risk-adjusted financial returns in a manner consistent with the investment policy. Investment decisions made for reasons other than the inherent economic and financial merit (that is, based in part on social or political concerns) should be set out in the investment policy and publicly disclosed. 20. The fund should not seek or take advantage of privileged information of exert inappropriate influence on private sector or other government entities. 21. If the fund exercises ownership rights associated with its investments, it should endeavor to protect the value of its investments. It should also publicly disclose its general approach to voting securities of listed entitles. 22. The fund should maintain a system of internal controls that identifies, assesses and manages all risks associated with its operations. Its risk management framework should be publicly disclosed. 23. The fund’s assets and investment performance should be reported according to clearly defined principles or standards. 24. The fund should establish a process of regular review of the implementation of these principles. -84- Appendix 5: Simulation of Reform Options This appendix uses projections generated by the Pensions Reform Options Simulation Toolkit (PROST) to evaluate the cost implications and impact on benefits for multiple parametric and structural reform measures. A. Parametric Reforms of the NDC Scheme Table 9 summarizes the impact of the baseline and proposed reform scenarios. It compares the current balance of the baseline and all the reform scenarios projected. It also indicates the long-term current balance observed as a proportion of GDP in the years observed. 70 Finally, the table summarizes the projected long-term levels of income replacement provided by pension insurance for new male retirees as a proportion of the projected average covered wage in the years observed (replacement rates). The table points out that the baseline projection suggests a financing gap of over 900% of 2011 GDP which results in annual deficits of about 5.8% of annual GDP and provides a long-term replacement rate of about 34% for new male retirees. Removal of the minimum pension guarantee materially reduces the financing gap yet also materially reduces the long-term replacement rate. A combination of parametric reforms to the NDC scheme as suggested in Scenario 2E results in a financing gap of about 177% of 2011 GDP, annual deficits of about 1.3% of GDP in the long-term, and a long-term replacement rate of about 42% for new male retirees. Extending the current defined-benefit scheme to post-1960 cohorts (scenario 3A) generates an even larger financing gap of over 1,300% of 2011 GDP though also provides a replacement rate of about 54% over the long-term for new male retirees. When the same reform parameters as applied in the reform scenario above (scenario 2E) are applied to defined-benefit or points schemes (scenarios 3B and 3C), the result is a financing gap of about 260% of GDP, annual subsidies of about 1.2% of GDP and a long-term replacement rate of about 35% for new male retirees. The results of adopting a Funded Defined Contribution (FDC) scheme depend upon the assumed rate of return on pension assets (Scenarios 4a, 4b, 4c). If the rate of return on pension assets is the same as wage growth (Scenario 4b), the long-term replacement rate for new male retirees is about 41%, almost the same as under a modified NDC scheme (scenario 2E). The financing gap however is greater under this scenario when compared to the modified NDC scheme (Scenario 2E) because of fiscal burden of transition costs. 70 The current balance is calculated based on the total contributions, interest, dividends or notional interest minus disbursements and administrative costs. A negative current balance or current financing gap represents a loss that generally needs to be compensated by a State Subsidy or the depletion of pension reserves or both. -85- Table 9: Summary of Impact of Baseline and Proposed Reform Scenarios Scenario Description Financing Long- Long-Term Gap-Present Term Replacement Value of PAYG Current Rate for New Current Balance Male Balances (% of Retirees (% 2011-2083 GDP) of Average (% of 2011 Wage) GDP)* 1 Baseline -918% -5.8% 34% 2A Removal of minimum pension guarantee -190% -0.7% 18% 2B Increase in retirement age & removal of -190% -1.1% 26% minimum pension guarantee 2C Increase in contribution rate & removal of -78% -0.3% 24% minimum pension guarantee 2D Combination of “2A-2C� -43% -0.6% 36% 2E 2D & actuarially fair annuity factors -177% -1.3% 42% 3A Retains the accrual rate of the defined-benefit -1314% -8.7% 54% scheme, applied to pre-1960 cohorts (i.e. 2.25% per year for first 20 years then 1.5% per year thereafter) for post-1960 cohorts 3B Defined-benefit formula remains changed to -259% -1.2% 35% linear accrual rate. Wage base used for income averaging is gradually increased to a career average. Wages are valorized in relation to wage growth. Accrual rate is set at a level of 1% in order to maintain the long- term current balance deficiency within 1% of GDP. Changes applied to years of service from 2013 onwards; same formulae as for pre- 1960 cohorts applied to years prior to 2013. Parametric reforms as in “2D� also introduced. 3C Points system replaces NDC scheme with -259% -1.2% 35% point value set at level equivalent to 1% accrual rate, applied to years of service from 2013 onwards. The same formula as for pre- 1960 cohorts is applied to years prior to 2013. Parametric reforms as in “2D� also introduced. 4 Application of a Funded Defined Contribution scheme (FDC) from 2013. Parametric reforms as in “2D� also introduced. Contributions made by the post-1960 cohort are directed to a separate dedicated pension fund. Benefits achieved are determined at different levels of rate of return. Disability and survivorship benefits are paid out on a PAYG basis at pre-reform levels. 4a Rate of return is 2% below average wage -430% -1.0% 24% growth 4b Rate of return is 2% above average wage -263% 0.5% 73% growth 4c Rate of return is equal to average wage -363% -0.4% 41% growth *At 2% real discount rate. -86- B.. Parametric Reforms of the NDC Scheme The reform options evaluated in this section examine the impact of: a) Removing the minimum pension guarantee. b) Increase the retirement age to 65 for men and women so that workers have a stronger incentive to work longer and bear more of the burden for their retirement income replacement. c) Gradually reinstating a 19% contribution rate to improve benefit adequacy. d) Adopting an actuarially fair annuity factor for calculating pension benefits. e) A combination of these four measures. The first reform scenario (2A) would remove the minimum pension guarantee for post-1960 cohorts. As suggested in Figure 26, this measure would reduce to a minimum the need for a state subsidy to the NDC scheme which – if there is no reform – is estimated to be approximately 6% of GDP per year over the long run. By removing the minimum pension guarantee and replacing it with a pensions-tested Tier I Targeted Social Pension, the net fiscal cost would be substantially reduced and the fiscal expenditures would target those retirees roughly in the bottom half of the income distribution. Figure 26. Scenario 2A: Removal of the Minimum Pension Guarantee for Post-1960 Cohorts (annual current balance as a % of GDP) Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. Removing the minimum pension guarantee would also reduce the long-term replacement rate expected from the contributory scheme (Figure 27). The elimination of a minimum pension has a major effect on the replacement rates. In the long run, replacement rates for new male retirees would decrease from about 34% under the baseline to about 18% without the minimum guarantee. -87- Figure 27. Scenario 2A: Removal of the Minimum Pension Guarantee for Post-1960 Cohorts (replacement rates) Source: Bank estimates using the PROST model. Note: Average pension for new male retirees as percentage of the insured average wage for current contributors. The next reform scenario (2B) continues to exclude a minimum pension guarantee and introduces a gradual increase in the retirement age. The retirement age is projected to increase to 65 for both genders at a rate of 2 months per year for men and 4 months per year for women, i.e. the retirement age first increases in 2013 and reaches 65 by 2042. Early retirement is also gradually eliminated by 2042. Increasing the retirement age has an overall impact on the system finances similar to scenario 2A (Figure 28). It further improves the annual current balance in the short- to medium term as fewer people retire and more stay in the system and continue to pay contributions. But in the longer run – once higher retirement age translates into higher pensions – deficits slightly increase compared with scenario 2A. Raising the retirement age significantly increases NDC pensions as individuals both increase contributions and have shorter payout periods: replacement rates for new male retirees are estimated to increase almost 45% compared with NDC pensions in the baseline and 2A scenarios (Figure 29). 71 71 The effects which increase the replacement rate are twofold: (i) by retiring later, workers are assumed to work longer and therefore acquire larger NDC balances for retirement; and (ii) these balances are paid out over a shorter life expectancy at the higher retirement age. -88- Figure 28. Scenario 2B: Removal of the Minimum Pension and Increase in Ret. Age (annual current balance) 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 (1.0%) (2.0%) (3.0%) (4.0%) (5.0%) (6.0%) (7.0%) baseline reform 2a reform 2b Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. Figure 29. Scenario 2B: Removal of the Minimum Pension and Increase in Ret. Age (replacement rates) 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% .0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 pre-1960 post-1960-baseline reform 2a reform 2b Source: Bank estimates using the PROST model. Note: Average pension for new male retirees as percentage of the insured average wage for current contributors. Reform scenario 2C continues to exclude a minimum pension guarantee and introduces a gradual increase in the contribution rate back to the 2008 level of 19% of wages. 72 The current balance impact of increasing the contribution rate is substantially stronger compared with both scenarios 2A and 2B, despite the longer term effect it has on benefits once higher contribution rates result in higher NDC pensions (Figure 30). The long run impact on NDC replacement rates is almost the same as that of increasing retirement age in scenario 2B (Figure 31). 72 This is the combined employer and worker contribution rate. The increase is assumed to begin in 2013 and proceed at a rate of 0.5% per year over a 10 year period reaching 19% in 2022. -89- Figure 30. Scenario 2C: Removal of the Minimum Pension and Increase in Contribution Rate (annual current balance) 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 (1.0%) (2.0%) (3.0%) (4.0%) (5.0%) (6.0%) (7.0%) baseline reform 2a reform 2b reform 2c Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. Figure 31. Scenario 2C: Removal of the Minimum Pension and Increase in Contribution Rate (replacement rates) 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% .0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 pre-1960 post-1960-baseline reform 2a reform 2b reform 2c Source: Bank estimates using the PROST model. Note: Average pension for new male retirees as percentage of the insured average wage for current contributors. Reform scenario (2D) evaluates the combined effects of excluding a minimum pension guarantee, increasing the retirement age and the contribution rate. The positive impact on the current balance is the strongest among four scenarios. Surpluses accumulated in the medium term can be invested and investment income can be used to finance longer term deficits in which case the system may become financially sustainable (Figure 32). At the same time, higher retirement age and contribution rates make replacement rates even higher than in the baseline scenario in the long run (Figure 33). 73 73 The increase is slightly less for women at about 6-10% of average wages. -90- Figure 32. Scenario 2D: Removal of the Minimum Pension Guarantee, Increase in Ret. Age and Increase in Contribution Rates (annual current balance) 2.0% 1.0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 (1.0%) (2.0%) (3.0%) (4.0%) (5.0%) (6.0%) (7.0%) baseline reform 2a reform 2b reform 2c reform 2d Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. Figure 33. Scenario 2D: Removal of Minimum Pension, Increase in Ret. Age and Increase in Contribution Rates (replacement rates) 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% .0% 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 2011 pre-1960 post-1960-baseline reform 2a reform 2b reform 2c reform 2d Source: Bank estimates using the PROST model. Note: Average pension for new male retirees as percentage of the insured average wage for current contributors. Reform scenario (2E) evaluates the effects of applying actuarially fair annuity factors in the calculations of NDC pensions versus simple life expectancies. Using simple life expectancies assumes zero interest provided on the notional account balances after the pension has started to be paid. 74 If the same interest rate is granted to pensioners as to current contributors (the notional interest rate), the annuity factors will be lower than life expectancies for respective ages and genders and, hence, will generate higher pensions. This makes the system more expensive and moves the current balance curve downward by about 0.7-0.8 percentage points compared with scenario 2D (Figure 34). However, NDC 74 The benefit formula in the 1999 law suggests that the notional account balance is divided by the number of years of life expectancy at retirement and does not apply a notional interest rate to the remaining notional balance. -91- replacement rates increase to reasonably adequate levels of 40%-42% which is about the same as provided for pre-1960 cohorts. Figure 34. Scenario 2E: Removal of the Minimum Pension Guarantee, Increase in Ret. Age, Increase in Contribution Rates and Actuarially Fair Annuity Factors (annual current balance) 2.0% 1.0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 (1.0%) (2.0%) (3.0%) (4.0%) (5.0%) (6.0%) (7.0%) baseline reform 2a reform 2b reform 2c reform 2d reform 2e Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. Figure 35. Scenario 2E: Removal of Minimum Pension, Increase in Retirement Age, Increase in Contribution Rates and Actuarially Fair Annuity Factors (replacement rates) 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% .0% 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 2011 pre-1960 post-1960-baseline reform 2a reform 2b reform 2c reform 2d reform 2e Source: Bank estimates using the PROST model. Note: Average pension for new male retirees as percentage of the insured average wage for current contributors. C. Extension of a Defined-Benefit Scheme to Post-1960 Cohorts Scenario 3A applies the accrual rate currently applied to pre-1960 cohorts (i.e. 2.25% per year for first 20 years then 1.5% per year thereafter) for post-1960 cohorts. This effectively means the current NDC scheme is closed and the system goes back to the defined benefit scheme that existed prior to the 1999 reform. Extending the defined- -92- benefit scheme would ensure quite high replacement rate levels of above 50% (Figure 36). However, the cost of such scheme is projected to grow even faster than in the baseline scenario – to unsustainable levels of 9% to 10% of GDP in the long run (Figure 37). This and other issues with DB scheme were the reason why it was reformed in 1999. Figure 36. Scenario 3A: Reinstating Pre-1999 PAYG Defined Benefit Scheme (replacement rates) 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% .0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 pre-1960 cohorts post-1960 cohorts-baseline pre-1960 cohorts-reform 3a Source: Bank estimates using the PROST model. Note: Average pension for new male retirees as percentage of the insured average wage for current contributors. Figure 37. Scenario 3A: Reinstating Pre-1999 PAYG Defined Benefit Scheme (annual current balance) 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 (1.0%) (2.0%) (3.0%) (4.0%) (5.0%) (6.0%) (7.0%) (8.0%) (9.0%) (10.0%) baseline reform 3a Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. Under Scenario 3B, the current NDC scheme is converted to a DB scheme which at the same time undergoes parametric changes (Figure 38). The parametric reforms would include: • Elimination of the minimum pension guarantee from 2012 onwards (replaced by the Targeted Social Pension for some retirees) as in Scenario 2. -93- • Increase in the retirement age to 65 for men and women at a rate of 4 months per year women and 2 months per year men (i.e. first increase in 2013 and both gender retirement age is 65 in 2042) as in Scenarios 2B,D,E. • Increase contribution rate to 19% at 0.5% per year over a 10 year period, first increase in 2013 and reaching 19% in 2022 as in Scenarios 2C,D,E. • For post-1960 cohorts the annual accrual rate is set at 1% for years of contributions from 2013 onwards while same formulae as for post-1960 cohorts is applied for years prior to 2013. • Wage base averaging is gradually increased by one year every year from 2013 onwards until it becomes a career average. Wages are valorized in relation to wage growth. Figure 38. Scenarios 3B,C: Conversion to DB Scheme with Parametric Changes (annual current balance) 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 (1.0%) (2.0%) (3.0%) (4.0%) (5.0%) (6.0%) (7.0%) (8.0%) (9.0%) (10.0%) baseline reform 3a reform 3b/3c Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. Figure 39. Scenarios 3B,C: Conversion to DB Scheme with Parametric Changes (replacement rates) 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% .0% 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 2011 pre-1960 cohorts post-1960 cohorts-baseline post-1960 cohorts-reform 3a post-1960 cohorts-reform 3b/3c Source: Bank estimates using the PROST model. Note: Average pension for new male retirees as percentage of the insured average wage for current contributors. -94- Similar to the effects of Scenario 2, elimination of the minimum pension guarantee, higher contribution rate and lower system dependency ratio (due to retirement age increase) dramatically improve the system finances and maintain deficits within reasonable limits of 1%-2% of GDP (Figure 38). Still, Scenario 3B is in general more costly for the government than any of Scenario 2 variants, even with comparable (Scenario 2D) or higher (Scenario 2E) replacement rates (Figure 34). As the new lower accrual rate is introduced very gradually, replacement rates – which drop initially because of elimination of the minimum pension guarantee – reduce very slowly and converge to about the same levels as in the baseline scenario (Figure 39). Under Scenario 3C a point system is gradually introduced for post-1960 cohorts. As in scenario 3B, the new rules apply to years after 2013, while the old rules currently effective for pre-1960 cohorts are applied to years prior to 2013. If the target replacement rate for a full career worker contributing for 40 years is set at 40% of the average wage, the outcomes with respect to the system finances and expected benefits would be identical to scenario 3B (Figure 38 and Figure 39). D. Establishment of a Funded Defined Contribution Scheme for Post- 1960 Cohorts In Scenario 4, an FDC scheme is simulated to replace the NDC scheme for post- 1960 cohorts. Under this scenario, no minimum pension guarantee is provided for post- 1960 cohorts, the retirement age is gradually increased to age 65 for men and women and the contribution rate is gradually increased to 19%. Contributions made by the post-1960 cohort are directed to individual accounts in a separate pension fund. The rate of return on pension fund assets has been varied under four scenarios as follows: Rate of Return on Scenario Pension Fund Assets Scenario 4a Wage Growth less 2% Scenario 4b Wage Growth plus 2% Scenario 4c Wage Growth Since the investment risk in FDC schemes is fully borne by individuals, expected benefits are very sensitive to the rate of return the pension fund will be able to earn on its assets (Figure 40). With the rate of return equal to wage growth (Scenario 4C) the projected replacement rates are similar to those projected in the NDC Scheme with similar increases in the retirement age and the contribution rates (Scenario 2E). If financial markets perform well and assets are invested wisely (Scenario 4B), system members can get up to 70% replacement rates and higher. But if financial markets are not developed well enough and assets are poorly managed, benefits can become inadequate – well below 30% as in Scenario 4A. With contributions of post-1960 cohorts diverted to the FDC scheme, the pension system will have to continue paying transition costs -- pensions to existing pensioners, to currently working pre-1960 cohorts once they retire as well as partially finance -95- disability and survivorship benefits. 75 Right after the reform, deficits in the remaining PAYG scheme are projected to practically double – drop to 5% of GDP –as a result of the transition costs. In the longer run, once pre-1960 cohorts retire and eventually quit the system, the PAYG current balance is reduced to disability and survivorship programs (Figure 41). Figure 40. Scenarios 4A,B,C: Converting NDC Scheme to Fully Funded DC (replacement rates) 80% 70% 60% 50% 40% 30% 20% 10% 0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 wage growth -2% wage growth wage growth +2% Source: Bank estimates using the PROST model. Note: Average pension for new male retirees as percentage of the insured average wage for current contributors. Figure 41. Scenarios 4A,B,C: Converting NDC Scheme to Fully Funded DC (PAYG annual current balance) 1.0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 (1.0%) (2.0%) (3.0%) (4.0%) (5.0%) (6.0%) wage growth -2% wage growth wage growth +2% Source: Bank estimates using the PROST model. Note: Current balance as a percent of current year projected GDP. 75 In Scenario 4 disability and survivorship benefits are assumed to remain at pre-reform levels and paid out of the PAYG component. Individual account balances in the funded DC component of those who become disabled or die are turned over to the PAYG component which tops up benefits to the levels guaranteed by the law if the balances are not enough to finance them or keep the difference if they exceed the amount needed. -96- Assets in the FDC scheme are projected to grow rapidly as the scheme matures and can reach very high levels of 100% to 180% of GDP depending on the assumed rate of return (Figure 42). This is an indicator of the potential supply of capital by the pension fund in the financial markets. Yet such accumulations also pose substantial challenges to the regulation and absorptive capacity of Mongolia’s financial markets. Figure 42. Scenarios 4A,B,C: Converting NDC Scheme to FDC Scheme (reserves in the funded DC scheme) 200.0% 180.0% 160.0% 140.0% 120.0% 100.0% 80.0% 60.0% 40.0% 20.0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 wage growth -2% wage growth wage growth +2% Source: Bank estimates using the PROST model. Note: Funded DC scheme reserves as a percent of current year projected GDP. -97- Appendix 6: Actuarial Projection Methodology A. Introduction This appendix describes the process used to establish a baseline for the financial projections of the Mongolian pension schemes. Mongolia has a single contributory national pension scheme for its entire labor force. The formal labor force comprising civil government workers, armed forces personnel and contracted private sector workers are covered by the system on a mandatory basis. The remaining labor force may participate in the system on a voluntary basis. The benefits to the formal workers and voluntary participants are the same. The scheme segregates the labor force into two cohorts: those born prior to 1960 (pre-1960 cohorts) and those born in 1960 or later (post-1960 cohorts). The benefits available under current law depend on the cohort group of the worker. Those in the pre- 1960 cohort are eligible for DB pension benefits, the main parameters of which are outlined below. Those in the post-1960 cohort have notional account balances and are eligible for NDC pension benefits, the main parameters of which are also outlined below. Post-1960 cohorts will begin to retire from 2015 onwards (early retirements and deaths leading to pensions can occur prior to this) and some of the parameters (such as average life expectancy) have yet to be established. The financial projections of the pension system were made using the Pension Reform Options Simulation Toolkit (PROST) an Excel-based software developed by the World Bank and licensed to country client users. The Toolkit has been enhanced over the years since its creation and the version used for this analysis is PROST 2006. The Toolkit follows the methodology depicted in Figure 43 for carrying out the projections and arriving at the inflows and outflows of a pension system. The information made available by SIGO during the World Bank mission in Nov- Dec 2010 pertained to different time periods but was complete for year ending December 31, 2008. Thus, the simulation period runs from 2008 to 2083. The total projection years were set to 75 in an attempt to capture the full life cycle of those currently entering the pension system. -98- Figure 43. PROST Projection Methodology Population Contributors Beneficiaries Economy PF revenues PS parameters PF expenditures PF balance -------------------------------- Implicit pension debt (IPD) -------------------------------- Individual accounts B. Demographic Projections and Assumptions B1.Country Population The first step taken by PROST is to project the country population for which the following inputs are required: • Current population distribution by age and gender • Fertility rates by age • Current and projected mortality rates by gender and age • Current and projected net migration by gender and age UN population data and projections for Mongolia were used in the simulations. Fertility is assumed to decline slightly from the current level of about 2 births per adult female to 1.85 births over the next 15 years and stabilize at that level throughout the rest of the simulation period. Mortality at all ages is expected to decline over the projection period consistent with expectations in countries of similar levels of development. This results in projected life expectancy as shown in Table 10. -99- Table 10: Projected Life Expectancy at Various Ages by Gender 2008 2020 2040 2060 2080 Male At Birth 63.7 67.5 72.1 75.2 78.4 At Age 20 47.7 50.6 54.2 56.6 59.0 At Age 55 18.6 20.1 22.2 24.1 26.1 At Age 60 15.4 16.7 18.4 20.1 21.8 At Age 65 12.6 13.6 15.0 16.3 17.8 Female At Birth 70.2 72.8 77.0 80.2 83.3 At Age 20 53.4 55.6 58.9 61.4 63.9 At Age 55 22.4 23.7 26.0 28.1 30.2 At Age 60 18.8 19.9 21.9 23.8 25.7 At Age 65 15.5 16.3 18.1 19.7 21.3 Source: PROST projections based on UN World Population Prospects, 2010 revision. Based on the above assumptions, population is projected to change as shown in Figure 44 below. The population shows steady growth until it peaks in 2053-54 touching almost 3.5 million but then shows a slight decline to end at approximately 3.3 million as fertility stabilize at levels below replacement rate. Figure 44. Population Projections Source: PROST projections based on UN World Population Prospects, 2010 revision. -100- Figure 45. Population Dependency Rates Source: PROST Projections based on UN World Population Prospects, 2010. A significant trend is the steady increase in the population dependency ratio which starts at 49% in 2009 and rises to 87% by 2050 (see Figure 45) 76 The ratio continues to increase to 101% by 2083 but the rate of increase is less steep after 2050. More significant to the current analysis is the rapid increase in the old-age (retirement age and older population) ratio to working age group (15 to retirement age) which rises from 11% in 2009 to 53% by 2050 and reaches 68% by 2083. C. Economic Assumptions In order for the cash flows to reflect nominal values, PROST requires inputs of inflation assumptions. In addition, knowing the nominal cash-flows does not provide sufficient information for policy makers to determine if the pension system is sustainable or not. The cash-flows must be measured against an indicator that provides information allowing such an assessment and for this purpose it is essential to project flow of GDP. C.1 Projected GDP Growth The debt sustainability analysis (DSA) produced jointly by the staffs of the IMF and the World Bank was used as a basis for GDP growth assumptions. This analysis covered the period 2010 to 2030 and took into account all macro-economic factors especially the impact of the expected growth in the mining sector. The analysis produced specific estimates for GDP growth for the years 2010 to 2015 and average growth for the period 2016-2030. For the purpose of this analysis the real GDP growth assumption has been set in line with the estimates in the DSA. After 2030, the assumption is that the GDP growth rate will gradually stabilize at around 2% which ensures a stable labor share in GDP given the assumed long term productivity growth of 2% per year and employment projections discussed below. See Table 11 for details. 76 Population dependency ratio is the sum of population in 0-14 and above retirement age groups in relation to the 15-retirement age group population. -101- C.2 Productivity Growth Wages grow due to aging of population (older workers on average earn higher salaries than younger workers) and also due to normal productivity growth. Generally speaking, productivity is strongly correlated to overall economic growth. If economy is doing well than workers can be expected to achieve greater wage growth: this correlation may not be visible over short durations but will hold true over longer periods. Productivity growth is expected to remain high during a period of high GDP growth, i.e. during period of mineral wealth extraction. However, there is likely to be lag between GDP growth and wage growth. In the long run productivity is assumed to converge to a steady state level of 2% per year. The real wage growth resulting from the productivity growth assumptions is shown in Table 11. C.3 Projected Inflation Inflation is extremely difficult to predict and is generally highly volatile in the short-term. However, since it underlies all economic assumptions, i.e. both in-flows and outflows, the level assumed is not critical to analysis or the decisions being determined. In accordance with the above mentioned DSA projections, it was assumed that inflationary pressures will remain during the period of strong economic growth and in the longer run will gradually converge to a steady state of 5%. See Table 11 for details. Table 11: Assumed Real GDP, Inflation and Wage Growth 2010 2011 2012 2013 2014 2015 2016 2030 2050 2083 Real GDP growth rate, % 13.7% 8.0% 1.0% 22.9% 15.7% 9.0% 8.3% 4.9% 1.4% 1.7% Inflation, % 10.2% 16.4% 16.0% 15.4% 14.8% 14.2% 13.6% 5.0% 5.0% 5.0% Real average wage growth rate, % 5.4% 7.1% 5.6% 6.1% 6.6% 7.2% 6.9% 3.6% 2.0% 2.0% D. Labor Force Projections The next step in the PROST projection process is to determine the numbers in the labor force and active employment. The inputs required for projecting the labor force and active workers are as follows: D.1 Labor Participation Rates The crude labor force participation rates indicated in the MSWL Statistics 2010 were used in the PROST projections. These are 65.7% for males 62.6% for females. 77 There is no additional information on the age based distribution of the labor force. 77 The labor force participation rates in the MSWL Statistics 2010 appear to be worked out on the basis of male working age being 16-59 and for females being 16-54. -102- Table 12: Assumed Labor Participation Rates MALES FEMALES Projection Year Projection Year Age 2008 2050 2083 Age 2008 2050 2083 16-60 64.6% 74.3% 83.9% 16-55 61.9% 71.2% 80.5% 61 36.5% 42.0% 47.4% 56 35.6% 41.0% 46.3% 62 23.8% 27.4% 31.0% 57 27.1% 31.2% 35.2% 63 21.1% 24.3% 27.4% 58 21.6% 24.8% 28.0% 64 16.1% 18.5% 20.9% 59 19.0% 21.8% 24.6% 65 14.0% 16.1% 18.2% 60 11.9% 13.7% 15.5% The pension system contributor information provided by SIGO indicated a number of contributors above the normal retirement age for both men and women. To remove any conflict of having contributors but no labor force for these ages, participation rates were applied to the older ages assuming contributors represented 40% of workers in each age group. A reduction was applied of 1.13% for males and 0.66% for females to achieve the aggregate participation rates actually observed. The labor force participation rates are projected to increase as the economy grows and improved conditions trickle down to workers. The assumed participation rates for 2008, 2050 and 2083 are shown in Table 12. PROST determines the rates for intermediate years using linear interpolation. D2. Unemployment Rates It was assumed that unemployment rates, which are at par if not better than for most mature economies, will continue at their current levels for both genders. The rates are applied to all ages in the working age range. The unemployment rates indicated in the MSWL Statistics 2010 are 2.5% for males 3.7% for females. There is no additional information on the age based distribution of the unemployed. E. Pension System Contributors The information provided by SIGO during the Nov-Dec 2010 mission included individual contributor data but for only those that had individual accounts, i.e. the pre- 1960 cohort’s information was not included. Additional information was received from SIGO in February 2011 which provided individual data for the pre-1960 cohorts giving a complete picture of contributors for the year 2008. The total number of full year contributors in the base year (2008) was estimated at 497,000 persons. The age distribution of contributors was developed for each gender using individual records. There were approximately 1.2% of male contributors above age 65 and 1.4% of females above age 60. The ages of these contributors were considered suspect and they were removed from the active contributor distribution and remaining were adjusted upwards to achieve the same annual wage bill. PROST offers a number of alternative methods for projecting the number of contributors in the future. In this analysis the “Stock of Employment� option was used -103- under which for each age and gender group future changes in the number of contributors are linked to changes in the employed population for the respective age and gender cohorts. Over time, these ratios are assumed to increase slightly to reflect the trends in the pension system participation, so the total coverage rate gradually increases from the current 46% to about 54% of the labor force by the end of the projection period. The projected number of contributors is driven by demographic projections – total population growth as well as its age composition – and coverage rate assumptions. In the longer run slowing down population growth rates and lower fertility rates result in stabilization of the total number of contributors (Figure 46). F. Pension System Beneficiaries The next step in PROST projection process is to determine the numbers of different categories of beneficiaries. Future changes in the number of beneficiaries, again, can be modelled in a number of different ways. For this analysis the flow method was used for the population of old-age and disabled; survivor and orphans beneficiaries were modelled as “Stock of Population�. In the baseline scenario the existing retirement pattern with respect to the age composition of new old age and invalidity pensioners was assumed to remain the same throughout the projection period. Survivors and orphans were assumed to change proportionately to changes in the population in respective age and gender cohorts. F1. Old-Age Pensioners by Gender/Age Distribution According to SIGO data, the total number of old age pensioners in 2008 was 193,000. SIGO also provided individual information on all beneficiaries receiving benefits. This data was used to develop the age/gender distribution of old-age pensioners. It was noted that although the female contributors exceeded male contributors by 11%, the number of female old-age pensioners was almost double the number of male old-age pensioners. -104- Figure 46. Projected Number of Contributors and Beneficiaries 900.0 800.0 700.0 600.0 500.0 400.0 300.0 200.0 100.0 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 contributors old age pensioners total beneficiaries Source: PROST projections. F2. Disabled Pensioners by Gender/Age Distribution SIGO provided information which could identify individual disabled pension beneficiaries and was used to develop the age/gender distribution of disabled pensioners. This allowed the current stock of disabled pensioners to be entered by gender and age and also for the stock as a percentage of population to be determined. The total number of disabled in 2008 was 43.5 thousand persons. F3. Survivor and Orphan Pensioners by Gender/Age Distribution SIGO provided information also allowed individual survivor pension beneficiaries to be identified but did not specifically distinguish between adults and orphans. It was assumed that those aged 19 or above constituted survivors and the remainder (i.e. below age 19) formed the orphans group. The total number of survivors in 2008 was about 22 thousand persons. Due to growing life expectancy, the total number of old age pensioners – the largest beneficiary group – is projected to grow faster than the number of contributors resulting in growing system dependency ratio (Figure 47). G. Pension System Cash-Flows With the projection of the population, labor force, employed, contributors and beneficiaries complete, PROST proceeds to determine the cash in-flows and out-flows of the pension system. -105- Figure 47. Projected System Dependency Ratio (the number of old age pensioners as a proportion of the number of contributors) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 2011 Source: PROST projections. G.1 Pension System Cash In-Flows A pension system can have more than one source of in-flows: contributions from workers and employers, government contributions, investment income, etc. For this analysis, government is included as an employer but not as a separate source of financing. If the system runs a deficit and has no reserves the government finances the deficit from the general budget. To project contribution flows it is necessary to provide following additional PROST inputs as specified below. G1.1. Cumulative Distribution of Wages by Gender The percentage of workers earning the minimum wage and cumulative percentage for higher wage brackets is required for each gender. This distribution was derived from the contributors’ information provided by SIGO for the year 2008. There was a small fraction of workers (approximately 6.44%) for whom the monthly wage worked out to below the national minimum wage. This may be due to natural factors such as workers doing fractional work months and the database only recording whole months; for workers at the minimum wage their average monthly wage will be below the minimum national wage as it was enhanced to MNT 108,000 per month during 2008; etc. For the purpose of this analysis, all workers whose average monthly wage was below MNT 108,000 were assumed to be earning this wage at 31.12.2008. Based on contributor individual records the average annual wage of contributors in 2008 is estimated at about MNT3.3 million. The minimum wage in Mongolia has not been revised from 2008-2010 but was revised three times during 2007. It is expected that it will be enhanced from 2011 onwards in line with the average wage growth. G1.2 Earnings Profile The earnings profile by gender and age cohort as a multiple of minimum wage is also required PROST input. This profile was derived from the same SIGO contributors’ -106- database. The adjustment to workers with average wage below the national minimum was also made while determining the earnings profile. In accordance with the international trends, in Mongolia older ages tend to receive higher wages. G1.3 Contribution Rate The rate of contribution required from employee and employer is defined as a percentage of wages of the employee and is currently 14% on a combined basis for those covered under the pension system on mandatory basis. For those who register voluntarily the contribution rate is 9.5% of reported wage. PROST does not provide for segregation of workforce into different categories of contributors unless the distinction is based on wage level. A weighted average contribution rate – estimated at 13.9% – was used in the baseline scenario. PROST takes into account the compliance rate measuring the system ability to collect contributions. Base on the provided data the current compliance rate (or collection rate) is estimated at around 80%. Over time, it is assumed that administration of the system will improve and the collection rate will gradually increase to about 95% of contributions due to the system in each year. G1.4 Additional Income There can be other sources of income which may include penalties for late contributions, investment income on short-term surpluses, and other factors. Analysis for the period of 2007-09 of SIGO’s financials, as indicated by the SIGO Yearbook 2010, showed that additional income averaged 0.51% of contributions collected in that period. It was assumed that this level of additional income will continue throughout the projection period. G.2 Pension System Cash Out-Flows Out-flows from a pension system are benefit payments to beneficiaries and cost of administering the scheme (there can be others such as cost of managing investments but these are not relevant to this analysis). The following inputs are required for PROST to project these out-flows: G2.1 Old-Age Pension Distributions PROST requires two old-age pensioner distributions: first is a cumulative distribution showing percentage of pensioners at the minimum monthly pension and the accumulated percentage for higher pension brackets; the second is the average pension of each age cohort as a multiple of the minimum pension. Both of these distributions are input separately for each gender. According to provided data, 32% of men and 61% of women were receiving minimum pension in 2008. There was no minimum pension increase in 2009 and in 2010 it was raised to MNT105.3 thousand (monthly). Based on our discussions with the authorities, over the next 10 years minimum pension is assumed to grow in line with wages (which is the current trend) and after that – at the average rate of inflation and average wage growth. -107- G2.2 Years of Contributions at Retirement for New Old-Age Pensioners Analysis of the retiree database showed that the average years of contribution has experienced some decline in the past (see Table 13). However, it is expected that this decline will not continue into the future since it has shown signs of leveling out and the lengths of contributions are not excessive. For the purpose of this analysis, the average years of contribution experienced in 2010 by age of retirement were input into PROST and held unchanged for future years. Table 13: Average Years of Contribution for New Old-Age Retirees YEAR OF RETIREMENT GENDER 2006 & Earlier 2007 2008 2009 2010 Males 30.71 27.94 27.11 27.05 26.56 Females 25.25 25.29 24.49 24.47 24.46 The following factors determine change in the average pension for old age pensioners: (i) the benefit formula for new pension calculations for pre- and post-1960 cohorts; (ii) the assumed pension indexation policy; and (iii) changes in the age composition of old age pensioners. The projected average replacement rate for all existing old age pensioners under the baseline scenario is presented in Figure 48. Figure 48. Projected Average Replacement Rate for Existing Old Age Pensioners (percent of current year average wage) 40.0% 35.0% 30.0% 25.0% 20.0% 15.0% 10.0% 5.0% 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 2011 Source: PROST projections. G2.4 New Disability Pensioner Replacement Rates For disability pensioners PROST requires that the expected replacement rates be input separately for both genders and each age at which disablement may occur. Analysis of the years of contributions associated with disability retirees showed that they were relatively stable over the last four years (Table 14). The incidence of disabilities was also quite stable and relatively low with approximately 3,000 male and 2,350 female disablements per year. Thus, it was deemed appropriate to use the average years of contribution observed over this four year period to arrive at the expected replacement rates by gender and age for 2008 and assume that they remain unchanged for future years. -108- Table 14: Average Years of Contribution for New Disability Retirees YEAR OF RETIREMENT GENDER 2010 2009 2008 2007 Males 12.34 12.47 12.70 13.04 Females 11.20 11.33 10.95 11.06 G2.5 Survivor and Orphan Pensioner Out-flows To project out-flows arising from survivor pensioners, PROST requires the ratio of average new survivor pensions starting in 2008 (which were MNT907 thousand per annum) to average old-age pensions for 2008 (which were MNT 1,247 thousand per annum). Thus, the required ratio input was 73%. With respect to survivorship pensions for orphans, PROST requires ratio of average orphan pensions paid in 2008 (which were MNT 1,144 thousand per annum) in relation to average wage of contributors for 2008 (which were MNT 3,287 thousand per annum). This ratio comes to 35%. These inputs are sufficient for PROST to project the costs arising from the DB scheme applicable to the pre- 1960 cohorts (i.e. those born in 1959 or earlier). Under current regulations the post-1960 cohorts (i.e. born 1960 or later) will be entitled benefits under the NDC scheme introduced with effect from 2000. The additional inputs required to project the benefits for the post- 1960 cohorts under the NDC scheme are as follows: i. Initial Capital for Post-1960 Cohorts The way that NDC benefits for post-1960 cohort are projected is that PROST develops a “reform� scenario for these cohorts starting from the very first projection year of 2009. For this purpose, the accrued rights have to be defined for those “switching� to the NDC scheme which is done by transferring the accumulated balances to the notional accounts of the switchers, these are averaged on gender and age groups. The average accumulated balances are input as percentages of overall average wage of the contributors for 2008. ii. Post-1960 Cohorts Contributions and Interest Rate on Balances The rate of contribution is assumed to be the same under both the DB and NDC schemes. The interest rates credited on accumulated balances in the notional accounts are based on average of three years wage growth rates. The projected notional interest rates were worked out and their relationship determined to the annual wage growth, these rates were entered into PROST. iii. Post-1960 Cohorts Pension Calculation The rate of annual pension under the NDC scheme is determined by dividing the accumulated balance at retirement by an “Average Life-Expectancy Factor�. These factors have not as yet been explicitly defined in the Mongolian system. For the purpose of this analysis these factors have been determined based on the projected life expectancies for respective age and gender cohorts. iv. Post-1960 Cohorts Minimum Pension -109- The NDC scheme has a minimum pension for those retiring with 15 years of contributions equal to 20% of national average wage, plus 0.5% of national average wage for each additional year of contribution. As stated earlier the DB scheme has a 75% of minimum wage for full pension cases and 50% of minimum wage for partial pension cases. Table 15: Defined Benefit and NDC Scheme Minimum Pensions Pension Minimum Minimum Wage Average Wage Type Amount MNT108,000 pm MNT270,390 pm DB Full Pension Minimum MNT81,000 75.0% 30.0% DB Partial Pension Minimum MNT54,000 50.0% 20.0% NDC Pension Minimum - 15 Years MNT54,080 50.1% 20.0% NDC Pension Minimum - 20 Years MNT60,840 56.3% 22.5% NDC Pension Minimum - 25 Years MNT67,600 62.6% 25.0% NDC Pension Minimum - 30 Years MNT74,360 68.8% 27.5% Source: Calculated from prevailing minimum and average wage amounts and prevailing parameters at December 31, 2010. The change in minimum pension level from DB to NDC will depend on the years of service pattern at retirement under the two schemes (See Table 15): The current average years of contribution are 26.56 and 24.46 for males and females respectively. This would appear to suggest that very little change will occur for those entitled to partial pension but some drop in minimum would result for those entitled to full pension. For the purposes of this analysis the pattern of minimum pensions has been assumed to be the same for both the DB and NDC schemes. Baseline projections for the system revenues, expenditures and the annual current balance based on the above inputs are shown in Figure 49. Figure 49. Projected Baseline Revenues, Expenditures and Current Balance 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% 2011 2016 2021 2026 2031 2036 2041 2046 2051 2056 2061 2066 2071 2076 2081 -2.0% -4.0% -6.0% -8.0% revenue expenditures current balance Source: PROST projections. -110-