Integrating a random utility random opportunity labour supply model in MIDAS Belgium: presentation of on-going work

Download Report

Transcript Integrating a random utility random opportunity labour supply model in MIDAS Belgium: presentation of on-going work

Integrating a random utility random
opportunity labour supply model in
MIDAS Belgium:
presentation of on-going work
Gijs Dekkers,
Federal Planning Bureau
CESO, KU Leuven
CEPS/INSTEAD
André Decoster
CES, KU Leuven
Bart Capéau
CES, KU Leuven
Federaal Planbureau
Economische analyses en
vooruitzichten
European Meeting Of The INTERNATIONAL MICROSIMULATION ASSOCIATION, October 23-24th, 2014, MAASTRICHT
Integrating a random utility random opportunity
labour supply model in MIDAS Belgium
–
–
–
–
–
Current versions of MIDAS include simple, reduced-form behavioural equations
Not ideal for reform analysis
complicating factor: MIDAS is dynamic
Another complicating factor: alignment
This presentation reports on on-going work to introduce the “random utility–
random opportunity model” (a.k.a. RURO) in the dynamic-ageing microsimulation
model MIDAS of Belgium.
– Brief overview of this presentation
• A birds-eye view on RURO
• Simulation in LIAM2: a simple example of code
• Oh, static is static, and dynamic is dynamic, and never the twain shall meet.
Wage thrift
Stability
Alignment
• Some preliminary results
Federaal Planbureau
Economische analyses en vooruitzichten
RuRo-model
Oslo model
standard model
– choice of discrete h

choice of j: (h,w,k)
– h: uniform distr.

h: non uniform
– gross wage given

gross wage distrib.
– tax-benefit system

tax-benefit system
– functional form U(.)

functional form U(.)
– assumptions about stochastic part

assumptions about stochastic part
– => prob (h)

=> prob (h,w)
Federaal Planbureau
Economische analyses en vooruitzichten
RuRo-model
• probability:
• standard multinomial logit-model
(relative attractiviness of the choice)
• RuRo
• weighted by measure of ‘availability’
Federaal Planbureau
Economische analyses en vooruitzichten
RuRo-model
• Structural => empirical specifications
– preferences
– opportunities (job availability)
Federaal Planbureau
Economische analyses en vooruitzichten
RuRo-model
• preferences: Box-Cox
Federaal Planbureau
Economische analyses en vooruitzichten
RuRo-model
• job availability
– market versus non-market
– market subset
Federaal Planbureau
Economische analyses en vooruitzichten
RuRo-model
• coefficients for utility function
• coefficients for opportunities
– market versus non market (q0)
– hours (peaks): g2(h)
– wage distribution: g1(w)
Federaal Planbureau
Economische analyses en vooruitzichten
RuRo-model
Some quite very extremely preliminary estimation results
Leisure coefficients M/F in couples
exponent
constant
ln(age)
ln(age)^2
# children between 0 and 3
# children between 4 and 6
# children between 7 and 9
region WAL
region BXL
Educ LOW
Educ HIGH
Leisure coefficients single M/F
exponent
constant
ln(age)
ln(age)^2
# children between 0 and 3
# children between 4 and 6
# children between 7 and 9
region WAL
region BXL
Educ LOW
Educ HIGH
Wage equation M/F
Sigma (RMSE)
constant
potential experience
potential experience^2
Educ LOW
Educ HIGH
Federaal Planbureau
males
Coeff
SE
t-value
females
Coeff
SE
t-value
-7.178
35.345
-19.054
2.686
-0.059
0.047
-0.100
0.255
0.207
-0.294
-0.055
0.543
11.778
6.464
0.898
0.084
0.089
0.088
0.104
0.163
0.128
0.093
-13.23
3.00
-2.95
2.99
-0.70
0.52
-1.13
2.45
1.27
-2.30
-0.59
-1.845
205.964
-115.314
17.345
1.232
1.646
1.219
2.131
0.545
2.334
-3.085
0.451
51.100
28.543
4.030
0.516
0.546
0.552
0.708
1.019
1.184
0.708
-4.09
4.03
-4.04
4.30
2.39
3.02
2.21
3.01
0.53
1.97
-4.36
-3.118
70.790
-38.329
5.610
0.000
-1.001
-2.742
2.509
0.765
-0.692
-0.881
0.705
43.123
23.933
3.334
0.000
2.263
1.318
0.774
0.740
0.736
0.645
-4.42
1.64
-1.60
1.68
0.00
-0.44
-2.08
3.24
1.03
-0.94
-1.37
-1.113
323.745
-177.290
25.346
3.706
0.914
-1.377
2.853
-2.365
1.811
-2.682
0.611
78.769
43.301
6.014
1.609
1.184
1.055
1.047
1.127
1.430
1.046
-1.82
4.11
-4.09
4.21
2.30
0.77
-1.30
2.72
-2.10
1.27
-2.56
0.253
2.037
2.420
-3.666
-0.146
0.242
0.004
0.027
0.225
0.500
0.017
0.014
63.73
76.25
10.77
-7.33
-8.36
17.38
0.256
2.010
2.275
-3.449
-0.097
0.280
0.004
0.026
0.228
0.565
0.022
0.015
59.47
77.29
9.98
-6.10
-4.32
18.35
Economische analyses en vooruitzichten
RURO in MIDAS BE: A simple example of LIAM2
code
ad_earnings:
args: gender, age
code:
[...]
return: [...]
Function: generate earnings
ad_welfare:
args: income
code:
[...]
return: [...]
Function: generate welfare benefit
ad_unemployment:
args: entitlement conditions
code:
[...]
return: [...]
Function: generate unemployment benefit
utility_optimisation:
- i: 1
- max_u: 0
- utility_rndm: normal(0.0, 1.0) * 100
- while:
cond: (i < 200)
code:
- joboffer: [make a MC simulation]
- hours: if(joboffer, [make a random draw of discrete hours], 0)
- hourly_wage: if(joboffer, ad_earnings(gender, age), 0)
- incomeW: if(joboffer, hourly_wage * hours, ad_unemployment(...))
- welfare: ad_welfare(incomeW)
the individual gets a job offer?
200Does
iterations
Draw a number of hours (or not)
- leisure: 1 - hours / (168 * 52)
- utility: function of (incomeW + welfare, leisure, utility_rndm)
- max_u: max(max_u, utility)
- opt_hours: if(i == 0, hours, if(max_u == utility, hours, opt_hours))
- i: i + 1
Federaal Planbureau
utility
Optimal choice after i iterations
Take max(utility)
Economische analyses en vooruitzichten
RURO in MIDAS BE: MIDAS is dynamic
•
•
Wages increase with productivity
Social and fiscal parameters increase, but at a lower rate in the short and
middle run
•
This will cause the RURO model to keel over as simulated time goes by
Federaal Planbureau
Economische analyses en vooruitzichten
Complicating factor: MIDAS is dynamic
Starting dataset ± 2.2K2 individuals in 2002
Simulate job-offers i, t
Draw hours i
LABOUR MARKET
MODULE t
i= 1 to 200
MIDAS
t=2002 to 2060
DEMOGRAPHIC
MODULE t
Simulate earnings i, t=A*
RURO
Simulate alternative incomes i, t=A
Derive net income i, t=A
PENSION & BENEFITS
MODULE t
Derive utility i, t=A*
Select hours where U(i)=Max, t
CONTRIBUTIONS AND
TAXATION MODULE t
REDISTRIBUTION,
POVERTY, INEQUALITY
OTHER OUTPUT
Federaal Planbureau
A = year of estimation – currently 2007
* Stochastic components are constant over t
(exception is ‘joboffer’ and only the random
component of earnings changes with labour
market transitions).
Economische analyses en vooruitzichten
Complicating factor: alignment
•
It is of course sad, but MIDAS is being used in a policy-assessment
environment.
• Therefore, we use alignment by sorting to be able to
assess policy measures in conjunction with a semi-aggregate
model
(see Dekkers, Inagaki and Desmet, 2012)
•
Alignment includes:
–
–
–
–
–
–
•
•
Who works and who does not
Unemployment
Early retirement/CELS
Private and public sector employment
…
And all this to age, gender and period
Hence, heterogeneity in choice sets needs to be included in an alignment
procedure in simulation.
Who receives a job-offer at period t?
– ‘risk’ based on individual characteristics, using estimation results of RURO
– Aligned to gender, age and period
Federaal Planbureau
Economische analyses en vooruitzichten
Complicating factor: alignment
Foreach t = 2002 to 2060
MC simulation of inversion at i
Foreach i = 1 to 200
Logit simulation of ‘risk’ joboffer
J(i) at i, given working(t – 1)
If inversion at i
Joboffer(i)=inverse(joboffer(i-1))
Joboffer(i) = ALIGNMENT(age, gender, t)
If Joboffer(i)
simulation of hours h
Unemployment benefit if eligible at t
simulation of earnings at A
MAX=max(MAX,
utility(i)
Federaal Planbureau
Apply means-test for welfare at A
Add family benefits
Derive net total income at A
Derive utitlity(i)
Economische analyses en vooruitzichten
Some extremely preliminary simulation results
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
joboffer MALE
Federaal Planbureau
joboffer FEMALE
MALE
FEMALE
Economische analyses en vooruitzichten
Some extremely preliminary simulation results
0.7
0.65
0.6
0.55
0.5
0.45
0.74
0.4
0.72
variant FEMALE
basis FEMALE
0.7
0.68
0.66
0.64
0.62
0.6
1
3
5
7
9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39
variant 90% MALE
Federaal Planbureau
basisvariant MALE
Economische analyses en vooruitzichten
Some extremely preliminary simulation results
optimal hours
45
40
35
30
25
20
1
3
5
7
9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39
base variant MALE
base variant FEMALE
variant MALE
variant FEMALE
optimal hours
38
36
34
32
30
28
1
3
5
7
9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39
base variant BRUSS
base variant FLAND
base variant WALL
Federaal Planbureau
Economische analyses en vooruitzichten
Some extremely preliminary simulation results
optimal hours men
42.5
42
41.5
41
40.5
40
39.5
39
38.5
38
1
3
5
7
9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39
base variant edu LOW M
base variant edu M M
optimal hours women
base variant edu H M
40
35
30
25
20
15
10
5
0
1
3
5
7
9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39
base variant edu L F
base variant edu M F
base variant edu H F
Federaal Planbureau
Economische analyses en vooruitzichten
Integrating a random utility random opportunity
labour supply model in MIDAS Belgium
Thank you
Federaal Planbureau
Economische analyses en vooruitzichten
Assumptions and hypotheses of the
Study Committee on Ageing
Key demographic hypotheses
2007
2030
2050
2060
Fertility
1.81
1.76
1.76
1.77
Men
77.3
81.2
84.0
85.3
women
83.3
87.0
89.7
90.9
Life expectancy at birth
Key macro
hypotheses
Up to
2011
2011-2014
≥ 2015
Yearly productivity
0.01%
1.28%
1.50%
14.75 in 2014
Decreasing towards 8%
Unemployment rate
Social policy hypotheses
2009-2010
≥ 2015
Wage ceiling
Current legislation
1.25%
Minimum right per working year
1.25%
Welfare adjustment non-lump-sum benefits
Employed and self-employed
0.50%
Welfare adjustment of lump-sum benefits
1.00%
Federaal Planbureau
Economische analyses en vooruitzichten