Non-financial (Notional) Defined Contribution Schemes World Bank Core Course on Pensions Day 2, Tuesday, November 8, 2011 Robert Holzmann World Bank and University of Vienna
Road Map I. Motivation for NDC approach II. III. IV. The Basic Design of NDC NDC-Type Developments and Country Experiences with NDC reforms (Latvia, Poland, Sweden, and Italy) Key Design and Implementation Issues V. Concluding Remarks VI. Selected References
I. Motivation for NDC Approach Limited it reform options for unfunded d first pillar due to inherited implicit pension debt and hence the costs of transition absent or underdeveloped financial markets Low credibility and limited success of parametric reforms Promises of NDC to handle key reform pressures Basic design is simple, transparent and credible Capacity to handle fiscal pressure from demographic, economic and political shocks Ability to address social-economic changes (aging, divorces, female labor force participation, etc) Attractive features to deal with challenges and opportunities of globalization Promising experiences with NDC reforms in Sweden, Latvia, Poland and Italy
II. Basic Design of NDC The Beauty of Simplicity Generic NDC Individual Equilibrium Generic NDC Macroeconomic Equilibrium Simple but a Few Rules Need to be Respected
The Beauty of Simplicity NDC is individual account on Pay-As As-You-Go basis (or like an illiquid life-time cash balance plan) Participants (and/or their employers) pay contributions on earnings during their whole career Contributions are noted on an individual account Account grows with contributions and is credited with rate of return (but remains unfunded/payg) At retirement, the notional capital is converted into an annuity which takes account of remaining life-expectancyexpectancy 5
Generic NDC Individual Equilibrium Individual notional capital T K i, T c w i, t I t t 1 I I t T T 1 t 1 1. (1 t ) Converting capital into annuity P j, K j, 1 G LE, LE 6
Generic NDC Macroeconomic Equilibrium i Static equilibrium condition K t + P t = L t A t = FA t + PA t K t : notional capital of all active workers P t : present value of all benefits in disbursement L t : total liabilities, A t : total assets FA t : financial assets, PA t : Pay-As As-You-Go Asset Dynamic equilibrium condition * L 1 FA 1 r PA 1 w t t t Permissible notional interest rate * FAt L t r PAt L t w FA t PAt L t L t 7
A sustainable rate of return on contributions exists (case of Jordan) 4.0% Real return on Real return on contributions = contributions = 1% growth average covered wage 2.0% Cash-balance (% GDP) (2%) (4%) (6%) 2003 2013 2023 2033 2043 Real return on contributions = 3% Real Return on contributions = 5% (8%) Current situation (10%)
Basic Principle of Solvency in the NA system Liabilities = Assets
Liabilities = Assets Pensions in payment Individual accounts
Liabilities = Assets Pensions in payment Financial assets Individual accounts?
Liabilities = Assets Pensions in payment Individual accounts Financial assets PAY-AS-YOU-GO ASSET IMPLICIT TAX
What happens in the insolvent payas-you-go system? Financial assets PAYG Asset Could be <0 Liabilities = Unfunded Liabilities Or Tax Overhang
Pay-as-you-go assets can be largely negative 0-2 Morocco Dijibouti Fraction of GDP -4-6 -8 Iran Lebanon -10 Jordan -12
How to compute the notional interest rate? The sustainable notional interest rate is actually the allowable growth rate of liabilities: the rate used to index pensions and revalorize individual accounts. Liabilities = Assets
New assets Liabilities = Assets
New Liabilities New assets Liabilities = Assets From changes in life expectancy
Allowable change New Liabilities New assets Liabilities = Assets From changes in life expectancy
NDC is simple but a few rules need dto be respected, inter alia Choice of appropriate interest rate, life expectancy A buffer fund to handle short-term term (financial and economic) shocks A balancing mechanism to address long-term shocks A financing mechanism to handle inherited ited commitments (legacy costs) when moving from existing system 19
III. NDC-Type Developments and Country Experiences NDC movement across the world NDC Experiences 10+ years after (Italy, Latvia. Poland, and Sweden) 20
NDC movement across the world Countries with NDC-reformed schemes (late 1990s) Sweden, Latvia, Poland, Italy Recently legislated NDC schemes Norway (2009), Egypt (June 2010) Countries with NDC-inspired reformed systems Brazil, Kyrgyz Republic, Russia Countries with NDC-type systems (point systems) Germany and France; Croatia, Romania, Slovakia, Ukraine Under discussion/consideration in Belarus, Czech Republic, France, Greece, Hungary, Lebanon, Spain, China,? All recent reforms in OECD countries mimicked elements of NDC reform (life-time earnings, decrements/increments, adjustment 21 of pension to change in life expectancy etc)
Lessons from NDC reforms in Sweden, Poland, Latvia and Italy Overall, reforms went well. Reform delivers adequate replacement rate for average earners and are financially sustainable for next 50 years Reform requires extensive preparatory work at technical and political level, good communication with the public, an efficient administrative framework, and good rules to cope with economic and demographic changes All 4 countries applied different pathways to reform, i.e. there is no single way to do, or different paths lead to the same final destination Reforms in all countries were undertaken against the backdrop of crises and 22 the need to adjust.
Year of implementation and coverage in NDC system Year of implementation Italy Latvia Poland Sweden 1995 1996 1999 1999 Covered Pi Private sector Universal Employees and self- Universal workers Public employees Self-employed workers Several schemes run by two major public retirement institutions employed Exceptions Few independent plans with a small number of Farmers Judges and prosecutors Military (2001) workers, mainly professionals Police (2001) Prison wards (2001) Border guards (2001) Miners (2005) 23
Contribution rates for pensions as Contribution rate for retirement savings, of which: percentage of wages Italy Latvia Poland *) Sweden 33% (employees) 20% 19.52% 18.5% 20% (self-employed) 24% (atypical contracts) NDC 33% (employees) 14% from 2012 12.22% 16.0% contribution 20% (self-employed) 24% (atypical contracts) FDC Voluntary scheme 6% from 2012 73% 7.3% 25% 2.5% contribution Occupational Yes, initially low but Yes, but very low Yes, but very low Yes, high coverage and voluntary systems gradually rising (20.1% of labour force in 2008) coverage coverage 24
Cohorts covered by NDC scheme Italy Latvia Poland Sweden Coverage policy Oldest cohort with NCD First cohort in full NDC Workers with less than 18 years of All contributors are in the new system Mandatory for people younger than 50, with Mixed old-new pension formula for contributing (as of exception of those transition cohorts. Past 1996) are in the new who can retire before recalculated according system. Those with 2009 to the existing data more than 18 years of contributing can opt for the new system. For those working before 1995, pension calculated according to a mixed formula. Mandatory: 1960 1941 1949 1938 1978 New entrants (1981) New entrants (1981) 1953 25
Individual accounts build-up up Notional rate of return Inheritance gain Italy Latvia Pl Poland Sweden GDP growth covered wage bill covered wage bill Per-capita wage growth growth growth Increases general reserve Increases general reserve Increases general reserve Equally divided between survivors and added to their notional accounts Additional credits to notional account: Insured periods of unemployment Based on past Based on wage, up to total of 5 years unemployment benefit, only when eligible for benefit (around 12 months) Maternity and No, but more parental leave generous transformation coefficient for mothers Taking care for disabled child (income tested) Conscripted military service Insured periods of income loss due to sickness Disability Occupational sickness/injury Post-gymnasium education 26
Pensions formulae Italy Latvia Poland Sweden Denominator Life expectancy adjusted with an imputed rate of return, set at the level of 1.5% Life expectancy (unisex) at retirement age Life expectancy (unisex) at retirement age Life expectancy adjusted with an imputed rate of return, set at the level of 1.6% Post-retirement indexation Retirement age To prices To prices since 2011 Mixed price-wage indexation with at least 20 per cent of wages 65/60 (optional for women), individuals with 36 years of contributions can retire at 61 62/62, but early retirement at 60 for individuals with 30 and more years of contributions into force until 31 December, 2011. To prices plus discrepancy between real wage growth and 1.6% used to compute annuity & balancing when liabilities exceed assets. 65/60 65/65 but the minimum retirement age is 61 Minimum qualifying period 5 years 10 years None, but required for a minimum guarantee (25/20) Minimum pension guarantee Minimum i Social assistance Depends of years of In 2009: 675 PLN 213b 2.13 base amounts benefit level pension (around 25% of average wage net of the income tax) paid from age 65 contributions. For individuals without a minimum qualifying period state social maintenance benefit (45 LVL per month) per month (around 20% of average wage), The MB is indexed as other pensions. for single pensioners, 1.90 base amounts per person for married couples. (base amount = 42 800 SEK in 2009 (around 15% of the Financing State budget Social insurance contributions; State social 27 maintenance benefit from General budget General budget revenue on the top of accrued benefit average wage) State budget on the top of the benefit
Country lessons Expected outcomes on labor force participation of elderly seems to have been achieved while replacement rates in line with OECD countries NDC appears to have weathered the storm created by the deep recession of 2009, albeit it did not emerge completely unscathed Expenditure projections suggests effectiveness of long-term fiscal balancing approach 28
Estimated impact of pension reform on participation i rates in 2020, in pp 25 20 15 10 5 0 Sweden EU27 Italy Poland 15-64 15-71 55-64 29
Expected replacement rate 30
Projected pension expenditure Change in expenditure (in percent of GDP) Country 2007-20 2020-30 2030-40 2040-50 2050-60 2007-60 Italy 0.1 0.7 0.8 0.8 1.1 0.4 Latvia 0.3 0.7 0.3 0.3 0.7 0.4 Poland 1.8 0.3 0.2 0.1 0.3 2.8 Sweden 0.1 0.1 0.1 0.3 0.3 0.1 EU27 0.4 0.9 0.7 0.2 0.2 2.4 Decomposition of public pension expenditure (in percent of GDP) Country 2007 level Dependency ratio contribution Coverage ratio contribution Employment effect contribution Benefit ratio contribution Interaction effect 2060 level Italy 14.0 10.4 3.2 1.1 5.5 1.0 13.6 Latvia 5.4 5.7 1.6 0.2 3.9 0.4 5.1 Poland 11.6 13.4 6.3 1.0 7.1 1.8 8.8 Sweden 9.5 5.6 0.4 0.4 4.3 0.6 9.4 EU27 10.1 8.7 31 2.6 0.7 2.5 0.6 12.5
Potential imbalances and automatic balancing mechanisms SWEDEN ITALY POLAND LATVIA Probability that a disequilibrium emerges Rate of return during the accumulation phase per capita wages GDP contribution base (1) contribution base Transformation coefficient -demographic assumptions backward looking backward looking backward looking forward looking - rate-of-return assumptions 1.6% real 1.5% real 1 1 Indexation of benefits (beyond CPI) per capita wages - 1.6% no partial (20% of wages) no Existence of automatic balancing mechanisms if L A Automatic adjustments in the notional linterest rate yes yes yes yes Coefficient update -timing yearly every 3 years yearly yearly - automatic yes yes yes yes - variables considered changes in life expectancy changes in life expectancy changes in life expectancy changes in life expectancy Automatic adjustments in the annuity yes no no no Reserve fund (2) yes no yes no (1) the real rate of return cannot be negative. (2) we consider only those which are not invested only in government bonds.
Poland: GDP and wage growth, NDC and FDC returns in comparison 33
IV. Key Design and Implementation Issues Choice of the notional interest rate proxy Design of explicit balancing mechanism Addressing the legacy costs of reform Role and size of reserve funding Estimation of remaining i cohort life expectancy and mechanism of risk sharing 34
Choice of notional interest rate proxy Purpose: Critical for assuring solvency of system Approach: Direct estimation of notional interest rate via liability-asset asset projections or choosing proxies such as GDP, wage or contribution base growth Issues: The better the proxy, the less need for balancing mechanism intervention. ti Need to explore direct estimation approaches. Approaches: GDP growth (Italy), covered wage bill growth (Latvia and Poland), per-capita wage growth (Sweden) 35
How do alternative proxies perform? Wages revalorized by growth rate of average wage; Pensions indexed by prices. Source: Robalino-Bodor, 2008
Wage bill Source: Robalino-Bodor, 2008
Sweden Source: Robalino-Bodor, 2008
Based on Projection of Pay-as-you-go Asset Source: Robalino-Bodor, 2008
Design of explicit balancing mechanism Purpose: Assuring solvency of system Approach: Adjusting notional interest rate proxy and pension indexation rule in line with estimated asset-liability deviation. Issues: Little explored topic at conceptual and empirical level. Requires better understanding of direct interest t rate estimates t compared to proxies (i.e. asset and liability projections) and decision on triggers and phasing. Approaches: Only Sweden has explicit and direct approach. Others use discretionary adjustments of indexation. 40
Role and size of reserve funding Purpose: Assuring liquidity of system Approach: Creating reserves that are able to deal with short/medium-term term macro-shocks and/or long- term temporary demographic bulges Issues: Size depending on objectives and risk preferences (6-24 months or 60 months expenditure), and build-up up (easier in immature systems) Approaches: Sweden inherited it fund. Others have no explicit fund yet (some have wealth fund) 41
Addressing the legacy costs of reform Purpose: Starting new scheme with clean balance sheet; establishing credibility; avoiding distortions during transition to new equilibrium Approach: Defining and estimating the costs and finding extra-system financing (budget, privatization assets, coverage expansion) Issues: Size of the costs, and capability to find less distortionary ways of taxation (as wage tax) Approaches: Reduced NIR (Latvia, Poland), or ignored (i.e. budget financed) 42
China:. Phasing of Net Legacy Costs Under Different Degrees of Coverage Expansion (CE) and New Contribution Rates (15% - left - and 20% - right panel 80.0000 60.0000 70.00 50.00 60.00 percent of GDP) Legacy Costs Net (in 50.00 40.00 30.00 20.00 cent of GDP) Leg gacy Costs Net (in per 40.00 30.00 20.00 10.00 10.00 0.00 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 0.00 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 Years after Reform 10.00 Years after Reform LC N (CE: 0%) LC N (CE: 50%) LC N (CE: 100%) LC N (CE: 150%) LC N (CE: 200%) LC N (CE: 0%) LC N (CE: 50%) LC N (CE: 100%) LC N (CE: 150%) LC N (CE: 200%) Holzmann and Jousten, 2010 43
Estimation of remaining cohort life expectancy and mechanism of risk sharing Purpose: Assuring solvency and avoiding unplanned burden shifting (via adjustment mechanism) Approach: Searching for best method to project cohort life expectancy given current information Issue: No best practice yet emerged (given systemic underestimation in LE) and no consensus on how best to share risk for unexpected increase in life expectancy 44
V. Concluding Remarks NDC is a very promising approach to achieve sustainable, fair, and non-distortionary ti PAYG scheme (as part of multi-pillar approach) NDC is not fool proof. But done by the book offers less possibilities for political gaming Not all conceptual and operational issues have yet been solved. But they need to be addressed in any other system and reform. And they come out more directly in NDC schemes.
Selected References Chłoń-Domińczak, Agniszka, Daniele Frank and Edward Palmer. 2012. The First Wave of NDC Taking Stock Ten Years Plus Down the Road, forthcoming in Holzmann-Palmer and Robalino, eds. Holzmann, Robert, Edward Palmer and David Robalino, 2012, Non- Financial Defined Contribution (NDC) Systems: Progress and New Frontiers in a Changing Pension World, forthcoming. Holzmann, Robert and Alain Jousten. 2010. Addressing the Legacy Costs in an NDC Reform: Conceptualization, Measurement, Financing. IZA Discussion Papers, forthcoming in Holzmann-Palmer and Robalino, eds. Holzmann, Robert, and Edward Palmer, eds. 2006. Pension Reform: Issues and Prospects for Non-Financial i Defined Contribution ti (NDC) Schemes.. Washington, DC: World Bank. David Lindeman, David Robalino,and Michal Rutkowski. 2006. NDC Pension Schemes in Middle- and Low-Income Countries,, in op.cit. David Robalino and Andras Bodor. 2008. On the Financial Sustainability of the Pay-as-you-go Systems and the Role of Government Indexed Bonds. Journal of Pension Economics and Finance. David Robalino and Tatyana ana Bogomolova. 2006. Implicit Pension Debt in the Middle-East and North Africa: Magnitude and Fiscal Implications. Working Paper No. 46. June. Office of the Chief Economist for the MENA Region. www.worldbank.org/sp or pensions