Projecting Pensions and Retirement Age in France:

1
Projecting Pensions and


Retirement Age in France

• Some lessons from the Destinie I Model
• Didier Blanchet, Sylvie Le Minez
• Thanks to the collective work of
  the Destinies team
• INSEE, France
• IMA2007 Conference  Celebrating 50 Years of
  Microsimulation  
• Vienna, 20-22 August 2007

2
The DESTINIE I Model

• Main features and developments of Destinie
• Projecting the age at retirement what have we
  learned?
• Some results and some questions

3
Why microsimulation? because of the
complexity of the French system

• Multiplicity of regimes
• Private sector, Public sector, Self-employed
• Complexity of rules within regimes
• For example, the basic pension for wage earners
  in the private sector

4
P ?(A,Y) . Wref . Min(1,Y/Yref)

• ?(A,Y) proportion depending on Age at
  retirement and on the number of Years of
  contributions
• Wref reference wage
• Min(1,Y/Yref) prorating coefficient depending
  on the number of Years (1 if Y Yref)

5
P ?(A,Y) . Wref . Min(1,Y/Yref)
6
P ?(A,Y) . Wref . Min(1,Y/Yref)
7
P ?(A,Y) . Wref . Min(1,Y/Yref)
8
P ?(A,Y) . Wref . Min(1,Y/Yref)
9
P ?(A,Y) . Wref . Min(1,Y/Yref)
10
The general structure of the model
the basic structure

• A dynamic model
• Individual data derived from a household survey
  (the 1997 Financial Assets Survey)

• Demographic employment biographies
• A major exogenous variable the school leaving
  age
• Simple adjustments to fit macro projections

11

Adaptations to policy questions how was the
model developed?

• First version (1993) quite simple
• except for the treatment of kinship ties
• Known or computed for the initial year
• Updated through the simulation (until 2040) each
  year, the sample is renewed only with births,
  unions occurring within the sample

12

Policy questions and
model developments

• Stronger conditions for retirement, increased
  incentives to postpone retirement
• module simulating the choice of the age at
  retirement
• a better labor market module
• Extension of pension reform to public sector
• Costs of old age invalidity
• Survivors benefits, poverty of retired people

13
Age at retirement
which impact of the reforms?

14
Choice of the age at retirement

implementation in
the model

• Basic deterministic rule
• individuals retire when they are eligible to a
  full rate pension
• Behavioral rule based on an income-leisure
  trade-off à la Stock and Wise
• Estimation on French data few incentives before
  2003 reform ? ad hoc calibration

15
Some results
on mean age at retirement

• Young cohorts postpone more than the older ones

• Consequences much larger in the public sector in
  2003

16
Realistic projections in the long-run?

• Global simulated impact on the LFP rates (60-64
  years old) not that important in the long-run

• Destinie is more than a pure supply-side model

• but without feed-back on the labor demand

17
Realistic projections in the short-run?

• We do not observe such supply reactions in the
  short-run some effects are difficult to
  integrate in such a model and can be only
  transitory

• Improve the accuracy of the initial database and
  avoid inconsistencies between initial
  observations and projections

18
Perspectives

• What can we think about this supply model?
• Stock Wise only one option among others
• not so inaccurate
• How to deal with non financial determinants?
• complementary
  approaches
• What about Destinie ?
• Destinie I a useful model but more and more
  difficult to
• maintain
• Destinie II - New model under construction,
  more modular
• - A more complex and accurate
  labour market module