Labour market responses of survival pensioners: estimating a labour supply model and predicting the

1
Labour market responses of survival
pensionersestimating a labour supply model and
predicting the effect of a reformIMA 2007
Conference Celebrating 50 Years of
Microsimulation, 20-22 August 2007, Vienna

• André Decoster (CES) KULeuven
• Kristian Orsini (CES) KULeuven and IZA Bonn
• Guy Van Camp (CES) KULeuven and FPS Social
  Security Belgium

2
Aim of the exercise

• According to the current legislation
• net income of an individual may decrease if
• a survival pension is combined with labour income
• A proposal was made to tackle this activity trap
• What is the possible impact of the proposal on
  the labour supply of people with a survival
  pension?

3
Content of the presentation

• Exposition of
• the rules for combining a survival pension with
  labour market income in the current system and in
  the proposed reform
• Exposition of
• the instruments, necessary to evaluate this
  measure
• data that describe the affected group
• model used to evaluate the measure
• Evaluation of the measure

4
Survival pensions and labour income the current
system
Rules for wage earners without dependent children
With dependent children threshold of  18.189,93
and  20.918,42 All nominal amounts expressed in
prices of 2001
5
Survival pensions and labour income the current
system
For an individual with  10,000.00 survival
pension, a gross hourly wage of  13.00 and no
dependent children
6
Survival pensions and labour income the proposed
reform

• A test is executed on the sum of labour income
  and the survival pension
• If the sum is smaller than  23,231.07 (without
  dependent children) or  26,521.37 (with
  dependent children), the survival pension is not
  adapted
• If the sum is equal to or larger than this
  amount, the survival pension is reduced with 50
  of the amount above the threshold
• the new measure is only introduced for people
  with a survival pension but without a retirement
  pension

7
Survival pensions and labour income the proposed
reform
For an individual with  10,000.00 survival
pension, a gross hourly wage of  13.00 and no
dependent children
8
Survival pensions and labour income comparison
between the current system and the proposed reform
For an individual with  10,000.00 survival
pension, a gross hourly wage of  13.00 and no
dependent children
9
Evaluation of the measure which instruments do
we need?

• A sample with individual data
• information on survival pensions, retirement
  pensions and data relevant to determine labour
  supply (gross hourly wage, family composition, )
  
• A microsimulationmodel to compute net income
  before and after the reform
• A model that captures the choice between labour
  and leisure

10
Evaluation of the measure disposable data

• A sample of 305,019 individuals, non randomly
  sampled and weighting factors
• For each individual in the sample administrative
  data from a datawarehouse that covers labour
  market and social protection data.
• This dataset covers
• Labour market characteristics of wage earners
  including their wage
• Labour market characteristics of self employed
  including their annual income
• Information on several benefits like unemployment
  benefits, disability benefits and pensions
• data mainly for 2001, but also data on the labour
  market history of wage earners

11
Evaluation of the measure disposable data

• Identification of survival and retirement
  pensions according to Belgian law is possible
• Survival pensions can be generated in different
  regimes, we use the sum over all regimes
• We only observe paid survival pensions (i.e. not
  the right) but do impute missing values for
  certain cases
• A sample of 1,751 cases potentially affected by
  the measure (out of a sample of 305,019
  individuals)
• between 18-65,
• and available for the labour market as wage
  earner
• i.e. no student, not (pre) retired, sick or
  disabled,
• not active as self employed.

12
Evaluation of the measure computation of net
income before and after the reform

• The model MIMOSIS computes among other things
• social security contributions starting from a
  reconstructed gross income concept
• personal income taxes starting from a
  reconstructed taxable income concept
• social benefits based on a means test
• Gross income from labour on annual basis is
  determined as the product of gross hourly wage
  and number of hours worked per week times 52
• Determination of gross hourly wage
• If currently on the labour market based on
  observed gross income and performed number of
  hours
• If currently not on the labour market but before
  that on the labour market as wage earner then
  similar data from the past
• If none of these two cases, imputation of the
  minimum-wage ( 6.92)
• For the subset, potentially affected by the
  measure the latter had to be done in 62 of the
  cases
• With MIMOSIS the net income before and after the
  reform for hours per week ranging from 0 to 50
  that are multiples of 10

13
Evaluation of the measure computation of net
income before and after the reform
For an individual with  10,000.00 survival
pension, a gross hourly wage of  13.00 and no
dependent children
14
(No Transcript)
15
Increase in labour supply
16
(No Transcript)
17
Decrease in labour supply
18
Old aged people (65 or more)
Out of the labour market youngsters (until the
age of 18), students (from 18 to 25), (pre)
retired, disabled or old aged
65
Othe households
Single males
Population potentially active but labour supply
not modelled self employed and people living not
as single nor as couple
Single females
(Pre) pensioned
Sick or disabled
Self employed
Population in working age
Couples
Population potentially active of which labour
supply has been modelled employees, unemployed
or inactives living as single or in couple
25
Active young ones (18-25)
Students
18
Youngsters (until the age of 18)
0
19
Hypotheses of the labour supply model

• Perfect information
• Absence of adjustment costs (agents can change
  their labour supply strategy without any
  additional costs)
• Absence of dynamic adjustment (immediatly supply
  the appropriate number of hours)
• Optimality of the observed choice (no demand side
  or institutional constraints)
• Labour supply is static in the sense that no
  considerations on the long run are taken into
  account
• Stable preferences on short and medium term
• Survival pensioners have identical preferences as
  people living as single or in couple

20
Labour supply elasticities
21
Labour supply elasticities of the population
targeted by the reform
22
Effects of the reform on labour supply
23
Effects of the reform on labour supply

• Increase in hours supplied, mainly comes from
• middle aged females (45-55),
• with medium to low wages (but above the minimum
  wage)
• and medium to high household incomes (3th and 4th
  quintile)

Decreases in supplied hours are observed, but are
less pronounced than the increases. The overall
net effect is positive
24
Effects of the reform on the budget
25
Conclusions

• The reform is predicted to have
• positive effects on the labour supply of widows
  and widowers (830 FTE)
• a budgetary cost (11 mil EUR)
• but
• only labour supply has been modelled
• with administrative data
• and no sensitivity analysis has been carried out