1. On using dynamic microsimulation models to assess the consequences of the AWG projections and hypotheses on pension adequacy: simulation results for Belgium, Sweden and Hungary A report by Gijs Dekkers, Raphaël Desmet, Ádám Rézmovits, Olle Sundberg, Krisztián Tóth Presentation before the Working Group On Ageing Issues (SPC-WG-AGE) Brussels, March 27 th, 2015

2. On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy Initiative to use dynamic microsimulation models available in member states to simulate prospective developments in (pension) adequacy while taking into account the projections and hypotheses of the AWG as much as possible. A collaboration between Sweden, Hungary and Belgium Demonstrate the potential contribution of dynamic microsimulation Introduction and overview

3. Outline of this presentation Sustainability and adequacy: two sides of the same coin Indicators Microsimulation Results for Sweden, Hungary and Belgium Conclusions On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

4. Two sides of the same coin Financial and social sustainability SPC and AWG: a clear and recognized need for a joint assessment AWG: prospective indicators of sustainability and the benefit ratio. SPC: ‘current’ SILC-based indicators of adequacy (financial poverty risk, LWI, multidimensional poverty) and several prospective indicators (forward- looking theoretical replacement ratios based on theoretical cases). On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

5. The choice of indicators At-risk-of-poverty-rate Gini S80/S20 For the total population, and various subgroups Pensioners: those receiving a pension benefit Elderly (65+) Working population On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

6. Dynamic microsimulation models The characteristics that distinguish micro-simulation models from other models are that modelling is done at the level of the individual and the household, and that the point of departure of simulations is a survey or an administrative dataset representing actual individuals at a given point in time The dynamic element involves including a prospective element in simulations, either through reweighting or “updating each attribute for each micro-unit for each time-interval” Most models have a recursive (modular) structure Most models allow taking into account exogenous information On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

7. We require prospective dynamic microsimulation models, because… Prospective sustainability assessment should come with prospective (and consistent) adequacy assessment (AROP, Gini) Complementarity with Theoretical Replacement Ratios The EU2020 indicators (AROP, LWI) are defined on the household level Key elements in explaining the development of all types of incomes are the uprating regimes, i.e. the linkages between benefits and earnings. Ageing is a typically long-term and dynamic phenomenon Other arguments o Look beyond pre-determined groups: who wins, who loses? o How to model non-linearities and thresholds? o Many policy measures have a dynamic element On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

8. Dynamic microsimulation models Sweden - SESIM Ministry of Finance, the Economic Affairs Department Produces simulation results for the AWG Starting dataset is 300 K individuals in 1999 Includes financial wealth, home ownership, private savings The oldest and most developed model of the three used in this project Belgium – MIDAS Federal Planning Bureau Starting dataset is 300K individuals in 2001, expanded to 2.2 K 2 individuals Alignment Hungary – MIDAS_HU Central Administration of National Pension Insurance Developed on the basis of MIDAS Belgium Starting dataset is 2 K 2 individuals in 2012 The most recent model of the three On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

9. Results for Sweden Swedish pension system Reformed system implemented in 2003 Notionally Defined Contribution (PAYG/NDC) with a Defined Contribution system Pension-income-tested minimum benefit Applies to all those born 1938 or later Previous system: flat-rate pension, together with an earnings-related PAYG component The last cohorts with pension rights in the old system will retire around 2020 On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

10. Results for Sweden At risk of poverty rate On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

11. Results for Sweden Gini On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

12. Results for Sweden S80/S20 On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

13. Results for Hungary One-pillar, PAYG, DB pension system Benefit a function of contribution years, and average net income after 1987 Indexed to forecasted inflation On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

14. Results for Hungary At risk of poverty rate On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

15. Results for Hungary Gini On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

16. Results for Hungary S80/S20 On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

17. Results for Belgium Three pillars: Mandatory Statutory pension system, DB/PAYG + means-tested ‘Guaranteed Income for Elderly Persons Wage earners Civil servants Self-employed Retirement age 65, but retirement possible from 62 on, given minimum career length Equivalent periods Private occupational pension schemes Private individual savings and pension schemes Indexation regime On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

18. Results for Belgium At risk of poverty rate On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

19. Results for Belgium Gini On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

20. Results for Belgium S80/S20 On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

21. Results for Belgium: a quick word on the sensitivity tests AROP Gini On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy

22. Conclusions Financial and social sustainability: two sides of a coin SPC/ISG and AWG: a clear and recognized need for coordination This project shows the use of microsimulation in the joint assessment of adequacy and sustainability of pensions. It is remarkable that the results on adequacy complement the results on financial sustainability Sweden and Hungary: pension expenditures as a percentage of GDP decline; AROP increases (especially in Sweden). In Belgium, pension expenditures to GDP increases, but AROP decreases. On using dynamic microsimulation models to assess the consequences of the AWG projections and hypothesis on pension adequacy