The demographic impact on the German pension system and reform options Robert Fenge (University of Rostock, CESifo) Francois Peglow (MPI for Demographic Research, Rostock) Ausschuss für Sozialpolitik Jahrestagung, 30.9.-2.10. 2015 Bonn September 2015 Robert Fenge University of Rostock 1/30
I. Motivation Three factors determine demographic development: 1. Fertility 2. Mortality 3. Migration Which factor is most important for the increasing financial burden of the German pension system? September 2015 Robert Fenge University of Rostock Motivation 2/30
The true cause of the rapid ageing of the German population is the decline of fertility rates. (Sinn 2013) An increase in life expectancy in itself would imply that the population size is continuing to rise. This is not the case. Hence, low fertility is the main demographic problem. => child pensions Other economists emphasize the importance of public family support. A confusion of shrinking or ageing population? September 2015 Robert Fenge University of Rostock Motivation 3/30
II. The Pension Simulation Model - Population projections - Projections of labor force participation and employment - Modelling the budget of the German Statutory Pension System 1. Demographic Development - Fertility rates - Life expectancy (mortality rates) - Migration Data: 12 th coordinated population projection for Germany (Federal Statistical Office, 2009) September 2015 Robert Fenge University of Rostock The Model 4/30
Fertility Reference szenario: Total fertility rate increases until 2025 up to 1.6. Low fertility szenario: TFR rises to 1.4 High fertility szneario: TFR rises to 2.01 ( French case in 2010) Life Expectancy Reference szenario: Life expectancy at birth increases by 6.3 years (women; 89.2) and 7.3 years (men; 85) between 2010 and 2060. Low LE szenario: constant at levels of 2010: 82.9 (women) and 77.6 (men) High LE szenario: LE increases to 91.2 (women) and 87.7 (men) Migration Reference szenario: Net migration increases from 10 000 to 100 000 between 2010 and 2014 No NM szenario: NM is zero High NM szenario: linear increase to 200 000 until 2020; constant thereafter September 2015 Robert Fenge University of Rostock Demographics 5/30
2. Labor Market Future labor force participation rates: dynamic cohort approach (OECD: Burnieaux et al. 2004; Carone 2005) - considers changes in labor participation among different age and gender groups and projects cohort-specific trends into the future. - these trends can be traced back to social factors (e.g. longer schooling), demographic factors (e.g. lower fertility), institutional factors (e.g. reforms of early retirement laws) and economic factors (e.g. household income, part-time employment) September 2015 Robert Fenge University of Rostock Labor Market 6/30
Projection of Participation Rate Data: German Microcensus of the German Federal Statistical Office (2013a) Mean of gender-specific labor force entry,, and exit rates,, for 2007-20010 are used to project future PR: (1): Projection until age 49 (2): Projection from age 50 Adjustment for (i) longer education of younger cohorts (ii) early and legal retirement age (iii) under-reporting of marginally employed (1) (2) September 2015 Robert Fenge University of Rostock Labor Market 7/30
Employment and Wages Reference unemployment rate: constant 7.7 % (2010) Low UE variant: decrease to 4 % in 2060 High UE variant: increase to 10 % in 2060 Employees subject to social insurance contributions: - Age-specific rates of self-employed and civil servants and non-working persons from Microcensus data; - Yearly averages of quarterly data on socially insured employees (Federal Employment Agency 2013) Data on average gross wage income subject to social insurance contributions (German statutory pension insurance 2012) - Linear wage growth (i) 1.5 % (ii) 2.5 % (iii) 3.5 % - Earning profiles from age-specific income data (Federal Statistical Office 2013) September 2015 Robert Fenge University of Rostock Labor Market 8/30
Number of Pensioners - Use participation rates (e.g. of self-employed and civil servants) - Consider old-age pensions, disability pensions and survivor pensions - Youngest age to be eligible for a pension is 50; youngest age for old-age pensions is 60 - Disability pensions: observed age-specific share of all pensioners (Deutsche Rentenversicherung 2013); take account of observed number of disability pensioners below age 50; starting in 2011 number assumed to be constant. - Survivor pensions: share of female or male population of age 60 or older (Deutsche Rentenversicherung 2013) September 2015 Robert Fenge University of Rostock Pensioners 9/30
3. The Pension Budget Revenues - Contributions by employed and unemployed persons (on average 48% of full contributions) - Federal subsidy: general and additional subsidy - Further revenues, e.g. from assets and refunds, appr. 0.5% of total revenues September 2015 Robert Fenge University of Rostock Pension Budget 10/30
Expenditure - Pension payments for old-age, disability and survivor pensions - Old-age pensions: early retirement possible at age 60; shifted by 2 years according to new LRA - Survivor pensions: large pensions (55% of regular pension) Other expenditure: - Health insurance contributions of pensioners: Data form DRV; starting 2011 constant 15.5% - Administrative costs - Medical rehabilitation Buffer: - 0.2 to 1.5 monthly payments of own expenditure (here: share of total expenditure (DRV, 2012), starting 2012 80%. September 2015 Robert Fenge University of Rostock Pension Budget 11/30
III. Projection Results Source: Deutsche Rentenversicherung (2012) and own calculations September 2015 Robert Fenge University of Rostock Projection Results 12/30
Source: Deutsche Rentenversicherung (2012) and own calculations September 2015 Robert Fenge University of Rostock Projection Results 13/30
Source: Deutsche Rentenversicherung (2012) and own calculations September 2015 Robert Fenge University of Rostock Projection Results 14/30
Sensitivity 1. Wage growth September 2015 Robert Fenge University of Rostock Projection Results 15/30
2. Unemployment September 2015 Robert Fenge University of Rostock Projection Results 16/30
3. Mortality September 2015 Robert Fenge University of Rostock Projection Results 17/30
4. Fertility September 2015 Robert Fenge University of Rostock Projection Results 18/30
5. Migration September 2015 Robert Fenge University of Rostock Projection Results 19/30
IV. Decomposition of Demographic Factors Benchmark: Hold two factors constant to isolate the impact of the third factor 1. Total effect: The development of the contribution rate 2. Structural effect: Mortality and fertility fixed at the level of 2010, no net migration RVB develops as a result of the given population structure in 2010 and future economic development. September 2015 Robert Fenge University of Rostock Decomposition 20/30
3. Mortality effect: Constant fertility rate in 2010, no migration 4. Fertility effect: Constant mortality, no net migration 5. Migration effect: Constant fertility and mortality 6. The residual effect comprises all multiple interactions (second order effects) of the demographic factors September 2015 Robert Fenge University of Rostock Decomposition 21/30
V. Decomposition Results September 2015 Robert Fenge University of Rostock Decomposition 22/30
September 2015 Robert Fenge University of Rostock Decomposition 23/30
September 2015 Robert Fenge University of Rostock Decomposition 24/30
VI. Reform Option - Systematic approach to adjust the legal retirement age to the development of life expectancy - Indexation for legal retirement ages that allows for a life-expectancy neutral contribution rate development - Enlarging effect of longer lives on the contribution rate of the GRV shall be outbalanced by longer working careers - Use the estimated `mortality effect on the contribution rate of the GRV to calculate target values of a `benchmark contribution rate' that would have evolved when life expectancy and underlying mortality would have remained constant at the level of 2010 September 2015 Robert Fenge University of Rostock Reform Option 25/30
Modelling - Iterative mechanism to approximate the influence of a rising retirement age on labor force participation rates - A rise of the legal retirement age by two years results in a change of the effective average age of retirement by 1.5 years (Werding, 2011, 2013) - Adjust the mechanisms of the actual pension legislation to modified legal retirement ages - In particular, the minimal retirement age increases in line with the legal retirement age. Also pension deductions for early retirement are adjusted linearly September 2015 Robert Fenge University of Rostock Reform Option 26/30
VI. Reform Option September 2015 Robert Fenge University of Rostock Reform Option 27/30
Paths of Legal Retirement Age September 2015 Robert Fenge University of Rostock Reform Option 28/30
September 2015 Robert Fenge University of Rostock Reform Option 29/30
VII. Conclusion 1. Higher wages and lower unemployment have little impact on pension parameters 2. Past population structure and mortality most important drivers of pension parameters 3. Dynamic adjustment of legal retirement age to compensate for the mortality effect September 2015 Robert Fenge University of Rostock Conclusion 30/30
Appendix September 2015 Robert Fenge University of Rostock Conclusion 31
Number of Pensioners Non-working persons Retired self-employed and civil servants Not-Contributors September 2015 Robert Fenge University of Rostock Pensioners 32
Pensioners Disability pensioners (d: age-specific rate of disability pensioners; German statutory pension insurance 2013) Old-age pensioners Widow s pensioners Widower s pensioners September 2015 Robert Fenge University of Rostock Pensioners 33
3. The Pension Budget Revenues - Contributions by employed and unemployed persons: w SV N, x t : Employees subject to social insurance BE : Gross income RVB : Contribution rate Contributions of unemployed equal on average 48 % of full contributions. September 2015 Robert Fenge University of Rostock Pension budget 34
Federal Subsidy General subsidy: vir RVB : virtual contribution rate without federal subsidy Simplification in our model: Additional federal subsidy (EB : Enhancement allowance) September 2015 Robert Fenge University of Rostock Pension budget 35
Expenditure EP, t c c : Earnings points for cohorts c at age x = t c arw : Actual pension value Disability pensions: Old-age pensions ( x min = 60, shifted by 2 years according to new LRA) RZF: adjustment factor for early retirement September 2015 Robert Fenge University of Rostock Pension budget 36
Expenditure Widow s pension 0.55 Widower s Pension 0.55 September 2015 Robert Fenge University of Rostock Pension budget 37
Further Expenditure Contributions to the health insurance of pensioners Administrative costs Medical rehabilitation September 2015 Robert Fenge University of Rostock Pension budget 38
Expenditure September 2015 Robert Fenge University of Rostock Pension budget 39
Expenditure September 2015 Robert Fenge University of Rostock Pension budget 40